Dual-Track Update System Implementation: user can choose update to a release or to dev (commits) track.

1. What compute_type=auto Does (CTranslate2 Documentation)

  auto = use the fastest computation type that is supported on this system and device

  CTranslate2 internally handles:
  - GPU capability detection (Compute Capability 6.1-8.x+)
  - CPU instruction set detection (AVX, AVX2, AVX512)
  - Automatic fallback for unsupported types
  - NEW: Blackwell/sm120 INT8 workaround (PR #1937)

  2. RTX 50XX (Blackwell) Issue - https://github.com/OpenNMT/CTranslate2/issues/1865

  | Problem    | int8 variants fail with CUBLAS_STATUS_NOT_SUPPORTED on sm120 GPUs     |
  |------------|-----------------------------------------------------------------------|
  | Root Cause | cuBLAS INT8 kernels fail when matrix dimensions aren't divisible by 4 |
  | Fix        | https://github.com/OpenNMT/CTranslate2/pull/1937 - Merged Dec 2025    |
  | Solution   | CTranslate2 now disables INT8 for sm120, auto-selects float16 instead |

  3. Current WhisperJAV Architecture (Problem)

  ┌─────────────────────────────────────────────────────────────────┐
  │ resolver_v3.py                                                  │
  │ _get_compute_type_for_device()                                  │
  │   CUDA → int8_float16  ← HARDCODED                            │
  │   CPU  → int8          ← HARDCODED                            │
  └──────────────────────────────┬──────────────────────────────────┘
                                 ▼
  ┌─────────────────────────────────────────────────────────────────┐
  │ faster_whisper_pro_asr.py (lines 62-73)                         │
  │ DUPLICATE LOGIC:                                                │
  │   if compute_type is None or auto:                            │
  │     CUDA → int8_float16  ← REDUNDANT                          │
  │     CPU  → int8          ← REDUNDANT                          │
  └──────────────────────────────┬──────────────────────────────────┘
                                 ▼
  ┌─────────────────────────────────────────────────────────────────┐
  │ CTranslate2 WhisperModel                                        │
  │   - Receives explicit compute_type                              │
  │   - Cannot apply internal optimizations                         │
  │   - RTX 50XX users get CUBLAS_STATUS_NOT_SUPPORTED              │
  └─────────────────────────────────────────────────────────────────┘

  Problem: WhisperJAV duplicates CTranslate2's logic and doesn't benefit from upstream fixes.

  4. Proposed Architecture (Solution)

  ┌─────────────────────────────────────────────────────────────────┐
  │ resolver_v3.py                                                  │
  │ _get_compute_type_for_device()                                  │
  │   CTranslate2 providers → auto  ← DELEGATE TO CTRANSLATE2     │
  │   PyTorch providers     → float16/float32 (unchanged)           │
  └──────────────────────────────┬──────────────────────────────────┘
                                 ▼
  ┌─────────────────────────────────────────────────────────────────┐
  │ faster_whisper_pro_asr.py                                       │
  │   Pass compute_type=auto directly to WhisperModel             │
  │   KEEP: VRAM exhaustion fallback (try int8 if OOM)              │
  └──────────────────────────────┬──────────────────────────────────┘
                                 ▼
  ┌─────────────────────────────────────────────────────────────────┐
  │ CTranslate2 WhisperModel                                        │
  │   - Detects GPU capability (sm120 → no INT8)                    │
  │   - Selects optimal type automatically                          │
  │   - RTX 50XX → float16 (works!)                                 │
  │   - RTX 40XX → int8_float16 (fastest)                           │
  │   - CPU      → int8 (fastest available)                         │
  └─────────────────────────────────────────────────────────────────┘

README.md

@@ -231,106 +231,160 @@ Whisper sometimes generates repeated text or phrases that weren't spoken. Whispe
 
 ## Installation
 
-### Windows Installer (Easiest)
+### Windows (Recommended)
 
-Download and run: **WhisperJAV-1.7.4-Windows-x86_64.exe**
+*Best for: Most users, beginners, and those who want a GUI.*
 
-This installs everything you need including Python and dependencies.
+1. **Download the Installer:**
+   **[Download WhisperJAV-1.7.5-Windows-x86_64.exe](https://github.com/meizhong986/WhisperJAV/releases/latest)**
+2. **Run the File:** Double-click the downloaded `.exe`.
+3. **Follow the Prompts:** The installer handles all dependencies (Python, FFmpeg, Git) automatically.
+4. **Launch:** Open "WhisperJAV" from your Desktop shortcut.
 
-### Upgrading from Previous Installer Versions
+> **Note:** The first launch may take a few minutes as it initializes the engine. GPU is auto-detected; CPU-only mode is used if no compatible GPU is found.
 
-If you installed v1.5.x or v1.6.x via the Windows installer:
+**Upgrading?** Just run the new installer. Your AI models (~3GB), settings, and cached downloads will be preserved.
 
-1. Download [upgrade_whisperjav.bat](https://github.com/meizhong986/whisperjav/raw/main/installer/upgrade_whisperjav.bat)
-2. Double-click to run
-3. Wait 1-2 minutes
+---
 
-This updates WhisperJAV without re-downloading PyTorch (~2.5GB) or your AI models (~3GB).
+### macOS (Apple Silicon & Intel)
 
-### Install from Source
+*Best for: M1/M2/M3/M4 users and Intel Mac users.*
 
-Requires Python 3.9-3.12, FFmpeg, and Git.
+The install script auto-detects your Mac architecture and handles PyTorch dependencies automatically.
 
-**Recommended: Use the install scripts** (handles dependency conflicts automatically, auto-detects GPU):
+**1. Install Prerequisites**
+
+```bash
+# Install Xcode Command Line Tools (required for GUI)
+xcode-select --install
+
+# Install Homebrew (if not installed)
+/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"
+
+# Install system tools
+brew install python@3.11 ffmpeg git
+```
+
+> **GUI Requirement:** The Xcode Command Line Tools are required to compile `pyobjc`, which enables the GUI. Without it, only CLI mode will work.
+
+**2. Install WhisperJAV**
+
+```bash
+git clone https://github.com/meizhong986/whisperjav.git
+cd whisperjav
+chmod +x installer/install_linux.sh
+
+# Run the installer (auto-detects Mac architecture)
+./installer/install_linux.sh
+```
+
+> **Intel Macs:** The script automatically uses CPU-only mode. Expect slower processing (5-10x) compared to Apple Silicon with MPS acceleration.
+
+---
+
+### Linux (Ubuntu/Debian/Fedora)
+
+*Best for: Servers, desktops with NVIDIA GPUs.*
+
+The install script auto-detects NVIDIA GPUs and installs the matching CUDA version.
+
+**1. Install System Dependencies**
+
+```bash
+# Debian / Ubuntu
+sudo apt-get update && sudo apt-get install -y python3-dev python3-pip build-essential ffmpeg libsndfile1 git
+
+# Fedora / RHEL
+sudo dnf install python3-devel gcc ffmpeg libsndfile git
+```
+
+**2. Install WhisperJAV**
+
+```bash
+git clone https://github.com/meizhong986/whisperjav.git
+cd whisperjav
+chmod +x installer/install_linux.sh
+
+# Standard Install (auto-detects GPU)
+./installer/install_linux.sh
+
+# Or force CPU-only (for servers without GPU)
+./installer/install_linux.sh --cpu-only
+```
+
+> **Performance:** A 2-hour video takes ~5-10 minutes on GPU vs ~30-60 minutes on CPU.
+
+---
+
+### Advanced / Developer
+
+*Best for: Contributors and Python experts.*
 
 <details>
-<summary><b>Windows</b></summary>
+<summary><b>Manual pip install</b></summary>
+
+> **Warning:** Manual `pip install` is risky due to dependency conflicts (NumPy 2.x vs SciPy). We strongly recommend using the scripts above.
+
+**1. Create Environment**
+
+```bash
+python -m venv whisperjav-env
+source whisperjav-env/bin/activate   # Linux/Mac
+# whisperjav-env\Scripts\activate    # Windows
+```
+
+**2. Install PyTorch First (Critical)**
+
+You must install PyTorch *before* the main package to ensure hardware acceleration works.
+
+- **NVIDIA GPU:** `pip install torch torchaudio --index-url https://download.pytorch.org/whl/cu124`
+- **Apple Silicon:** `pip install torch torchaudio`
+- **CPU only:** `pip install torch torchaudio --index-url https://download.pytorch.org/whl/cpu`
+
+**3. Install WhisperJAV**
+
+```bash
+pip install git+https://github.com/meizhong986/whisperjav.git@main
+```
+
+</details>
+
+<details>
+<summary><b>Editable / Dev install</b></summary>
+
+Use this if you plan to modify the code.
+
+```bash
+git clone https://github.com/meizhong986/whisperjav.git
+cd whisperjav
+
+# Windows
+installer\install_windows.bat --dev
+
+# Mac/Linux
+./installer/install_linux.sh --dev
+
+# Or manual
+pip install -e ".[dev]"
+```
+
+</details>
+
+<details>
+<summary><b>Windows source install</b></summary>
 
 ```batch
 git clone https://github.com/meizhong986/whisperjav.git
 cd whisperjav
-installer\install_windows.bat              # Auto-detects GPU and CUDA version
+installer\install_windows.bat              # Auto-detects GPU
 installer\install_windows.bat --cpu-only   # Force CPU only
 installer\install_windows.bat --cuda118    # Force CUDA 11.8
 installer\install_windows.bat --cuda124    # Force CUDA 12.4
-installer\install_windows.bat --minimal    # Minimal install (no speech enhancement)
-installer\install_windows.bat --dev        # Development/editable install
-```
-
-The script automatically:
-- Detects your NVIDIA GPU and selects optimal CUDA version
-- Falls back to CPU-only if no GPU found
-- Checks for WebView2 runtime (required for GUI)
-- Logs installation to `install_log_windows.txt`
-- Retries failed downloads up to 3 times
-
-</details>
-
-<details>
-<summary><b>Linux / macOS</b></summary>
-
-```bash
-# Install system dependencies first (Linux only)
-# Debian/Ubuntu:
-sudo apt-get install -y python3-dev build-essential ffmpeg libsndfile1
-
-# Fedora/RHEL:
-sudo dnf install python3-devel gcc ffmpeg libsndfile
-
-git clone https://github.com/meizhong986/whisperjav.git
-cd whisperjav
-chmod +x installer/install_linux.sh
-./installer/install_linux.sh               # Auto-detects GPU
-./installer/install_linux.sh --cpu-only    # Force CPU only
-./installer/install_linux.sh --minimal     # Minimal install
 ```
 
 </details>
 
-<details>
-<summary><b>Cross-Platform Python Script</b></summary>
-
-```bash
-git clone https://github.com/meizhong986/whisperjav.git
-cd whisperjav
-python install.py              # Auto-detects GPU, defaults to CUDA 12.1
-python install.py --cpu-only   # CPU only
-python install.py --cuda118    # CUDA 11.8
-python install.py --cuda121    # CUDA 12.1
-python install.py --cuda124    # CUDA 12.4
-python install.py --minimal    # Minimal install (no speech enhancement)
-python install.py --dev        # Development/editable install
-```
-
-</details>
-
-**Alternative: Manual pip install** (may encounter dependency conflicts):
-
-```bash
-# Install PyTorch with GPU support first (NVIDIA example)
-pip install torch torchaudio --index-url https://download.pytorch.org/whl/cu124
-
-# Then install WhisperJAV
-pip install git+https://github.com/meizhong986/whisperjav.git@main
-```
-
-**Platform Notes:**
-- **Apple Silicon (M1/M2/M3/M4)**: Just `pip install torch torchaudio` - MPS acceleration works automatically
-- **AMD GPU (ROCm)**: Experimental. Use `--mode balanced` for best compatibility
-- **CPU only**: Works but slow. Use `--accept-cpu-mode` to skip the GPU warning
-- **Linux server (no GPU)**: The install scripts auto-detect and switch to CPU-only
-- **Linux (Debian/Ubuntu)**: Install system dependencies first: `sudo apt-get install -y python3-dev build-essential ffmpeg libsndfile1`
-
 ### Prerequisites
 
 - **Python 3.9-3.12** (3.13+ not compatible with openai-whisper)

RELEASE_NOTES_v1.7.5.md

@@ -0,0 +1,281 @@
+# WhisperJAV v1.7.5 Release Notes
+
+**Release Date:** January 2026
+
+---
+
+## New Features
+
+### In-App Upgrade System (Experimental)
+
+A complete upgrade system that checks for new versions and allows one-click updates.
+
+| Interface | Command |
+|-----------|---------|
+| **GUI** | Notification appears on startup when updates are available |
+| **CLI** | `whisperjav-upgrade --check` |
+| **CLI** | `whisperjav-upgrade --yes` (auto-confirm) |
+| **CLI** | `whisperjav-upgrade --wheel-only` (skip dependencies) |
+
+**Update Notification Intervals:**
+- Critical/security updates: Always shown
+- Major updates: Shown every launch
+- Minor updates: Shown monthly
+- Patch updates: Shown weekly (dismissable)
+
+> **Note:** This feature is experimental. Backup important files before upgrading.
+
+---
+
+### Resumable Processing
+
+| Feature | Description |
+|---------|-------------|
+| **Translation Resume** | Interrupted translations continue from where they stopped |
+| **Batch Skip** | `--skip-existing` flag skips files that already have subtitles |
+
+---
+
+### GPU Auto-Detection
+
+Installer and runtime now automatically detect GPU availability and fall back to CPU-only mode when no compatible GPU is found.
+
+---
+
+### Installation Scripts
+
+New standalone scripts for source builds with automatic CUDA detection:
+
+| Script | Platform | CUDA Support |
+|--------|----------|--------------|
+| `install_windows.bat` | Windows | Auto-detect (11.8, 12.1, 12.4, 12.6, 12.8) |
+| `install_linux.sh` | Linux | Auto-detect (11.8, 12.1, 12.4, 12.6, 12.8) |
+
+---
+
+## Bug Fixes
+
+| Issue | Description |
+|-------|-------------|
+| [#97](https://github.com/meizhong986/WhisperJAV/issues/97) | Fixed FFmpeg path not found in installer |
+| [#96](https://github.com/meizhong986/WhisperJAV/issues/96) | Added GUI documentation for tab relationships and settings persistence |
+| [#95](https://github.com/meizhong986/WhisperJAV/issues/95) | Fixed CUDA driver compatibility - falls back to CPU if incompatible |
+| [#94](https://github.com/meizhong986/WhisperJAV/issues/94) | Fixed Semantic options in Advanced tab |
+| [#92](https://github.com/meizhong986/WhisperJAV/issues/92) | Fixed version number display in GUI |
+| - | Fixed installer shortcut not launching application |
+
+---
+
+## Installer Improvements
+
+- Improved Windows installer robustness and completeness
+- Added `fsspec>=2025.3.0` constraint to prevent unnecessary downgrades
+- Added progress indication during pip install phase
+- NSIS now handles all shortcut creation (proper separation of concerns)
+- Added support for CUDA 12.8 (driver 570+)
+- Fixed desktop shortcut pointing to correct executable (`WhisperJAV-GUI.exe`)
+
+---
+
+## Documentation Updates
+
+- GUI documentation for tab relationships and settings persistence
+- Speech enhancement CLI documentation
+- Updated installation section with new scripts and features
+
+---
+
+## Installation
+
+### Upgrading from v1.7.4?
+
+If you are upgrading from version 1.7.4, simply run the new installer.
+
+- **Do not uninstall manually.** The installer will detect your previous version.
+- **Data Safety:** Your AI models (~3GB), settings, and cached downloads will be **preserved**.
+- **Clean Install:** The installer will automatically clean up old dependency conflicts.
+
+---
+
+### Windows (Recommended)
+
+*Best for: Most users, beginners, and those who want a GUI.*
+
+1. **Download the Installer:**
+   **[Download WhisperJAV-1.7.5-Windows-x86_64.exe](https://github.com/meizhong986/WhisperJAV/releases/tag/v1.7.5)**
+2. **Run the File:** Double-click the downloaded `.exe`.
+3. **Follow the Prompts:** The installer handles all dependencies (Python, FFmpeg, Git) automatically.
+4. **Launch:** Open "WhisperJAV" from your Desktop shortcut.
+
+> **Note:** The first launch may take a few minutes as it initializes the engine. GPU is auto-detected; CPU-only mode is used if no compatible GPU is found.
+
+---
+
+### macOS (Apple Silicon & Intel)
+
+*Best for: M1/M2/M3/M4 users and Intel Mac users.*
+
+The install script auto-detects your Mac architecture and handles PyTorch dependencies automatically.
+
+**1. Install Prerequisites**
+
+```bash
+# Install Xcode Command Line Tools (required for GUI)
+xcode-select --install
+
+# Install Homebrew (if not installed)
+/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"
+
+# Install system tools
+brew install python@3.11 ffmpeg git
+```
+
+> **GUI Requirement:** The Xcode Command Line Tools are required to compile `pyobjc`, which enables the GUI. Without it, only CLI mode will work.
+
+**2. Install WhisperJAV**
+
+```bash
+git clone https://github.com/meizhong986/whisperjav.git
+cd whisperjav
+chmod +x installer/install_linux.sh
+
+# Run the installer (auto-detects Mac architecture)
+./installer/install_linux.sh
+```
+
+> **Intel Macs:** The script automatically uses CPU-only mode. Expect slower processing (5-10x) compared to Apple Silicon with MPS acceleration.
+
+---
+
+### Linux (Ubuntu/Debian/Fedora)
+
+*Best for: Servers, desktops with NVIDIA GPUs.*
+
+The install script auto-detects NVIDIA GPUs and installs the matching CUDA version.
+
+**1. Install System Dependencies**
+
+```bash
+# Debian / Ubuntu
+sudo apt-get update && sudo apt-get install -y python3-dev python3-pip build-essential ffmpeg libsndfile1 git
+
+# Fedora / RHEL
+sudo dnf install python3-devel gcc ffmpeg libsndfile git
+```
+
+**2. Install WhisperJAV**
+
+```bash
+git clone https://github.com/meizhong986/whisperjav.git
+cd whisperjav
+chmod +x installer/install_linux.sh
+
+# Standard Install (auto-detects GPU)
+./installer/install_linux.sh
+
+# Or force CPU-only (for servers without GPU)
+./installer/install_linux.sh --cpu-only
+```
+
+> **Performance:** A 2-hour video takes ~5-10 minutes on GPU vs ~30-60 minutes on CPU.
+
+---
+
+### Advanced / Developer
+
+*Best for: Contributors and Python experts.*
+
+<details>
+<summary><b>Manual pip install</b></summary>
+
+> **Warning:** Manual `pip install` is risky due to dependency conflicts (NumPy 2.x vs SciPy). We strongly recommend using the scripts above.
+
+**1. Create Environment**
+
+```bash
+python -m venv whisperjav-env
+source whisperjav-env/bin/activate   # Linux/Mac
+# whisperjav-env\Scripts\activate    # Windows
+```
+
+**2. Install PyTorch First (Critical)**
+
+You must install PyTorch *before* the main package to ensure hardware acceleration works.
+
+- **NVIDIA GPU:** `pip install torch torchaudio --index-url https://download.pytorch.org/whl/cu124`
+- **Apple Silicon:** `pip install torch torchaudio`
+- **CPU only:** `pip install torch torchaudio --index-url https://download.pytorch.org/whl/cpu`
+
+**3. Install WhisperJAV**
+
+```bash
+pip install git+https://github.com/meizhong986/whisperjav.git@main
+```
+
+</details>
+
+<details>
+<summary><b>Editable / Dev install</b></summary>
+
+Use this if you plan to modify the code.
+
+```bash
+git clone https://github.com/meizhong986/whisperjav.git
+cd whisperjav
+
+# Windows
+installer\install_windows.bat --dev
+
+# Mac/Linux
+./installer/install_linux.sh --dev
+
+# Or manual
+pip install -e ".[dev]"
+```
+
+</details>
+
+<details>
+<summary><b>Windows source install</b></summary>
+
+```batch
+git clone https://github.com/meizhong986/whisperjav.git
+cd whisperjav
+installer\install_windows.bat              # Auto-detects GPU
+installer\install_windows.bat --cpu-only   # Force CPU only
+installer\install_windows.bat --cuda118    # Force CUDA 11.8
+installer\install_windows.bat --cuda124    # Force CUDA 12.4
+```
+
+</details>
+
+---
+
+## Technical Details
+
+### New Files
+- `whisperjav/version_checker.py` - Version checking and update notifications
+- `whisperjav/upgrade.py` - Self-upgrade CLI tool
+- `tests/test_upgrade_version_checker.py` - 44 tests for version checker
+- `tests/test_upgrade_cli.py` - 36 tests for upgrade CLI
+
+### Test Coverage
+- 80 automated tests for upgrade system
+
+---
+
+## Known Issues
+
+- Self-upgrade on Windows may require Administrator privileges if installed in Program Files
+- Update notifications require GUI restart after upgrade
+
+---
+
+## Contributors
+
+- MeiZhong - Development and testing
+- Claude (Anthropic) - Code assistance and documentation
+
+---
+
+**Questions or Issues?** [Open a GitHub issue](https://github.com/meizhong986/WhisperJAV/issues)

installer/generated/README_INSTALLER_v1.7.5.txt

@@ -0,0 +1,409 @@
+===============================================================================
+                            WhisperJAV v1.7.5
+                Japanese AV Subtitle Generator with AI
+===============================================================================
+
+Thank you for installing WhisperJAV!
+
+-------------------------------------------------------------------------------
+QUICK START
+-------------------------------------------------------------------------------
+1. Double-click the "WhisperJAV v1.7.5" desktop icon
+2. Select your video/audio files using "Add File(s)" or "Add Folder"
+3. Choose processing mode (Balanced recommended for best quality)
+4. Click "Start" to begin processing
+5. Subtitles will be saved next to your media files in an output folder
+
+FIRST RUN NOTE: On your first transcription, AI models will download (~3GB).
+                This is a one-time download that takes 5-10 minutes depending
+                on your internet speed. Progress is shown in the GUI.
+
+-------------------------------------------------------------------------------
+WHAT'S NEW IN v1.7.5 (HOTFIX RELEASE)
+-------------------------------------------------------------------------------
+Version 1.7.5 is a hotfix release addressing critical issues:
+
+1. CRITICAL BUG FIXES
+   - Fixed missing FFmpeg issue in conda-constructor installer
+   - FFmpeg now correctly bundled and accessible in PATH
+
+2. NEW EXPERIMENTAL FEATURES
+   - Self-upgrade system (EXPERIMENTAL)
+     * Check for updates via GUI or CLI: whisperjav-upgrade --check
+     * Automatic upgrade: whisperjav-upgrade --yes
+     * Note: This feature is in evaluation phase - use with caution
+   - Update notifications in GUI when new versions are available
+
+IMPORTANT: The upgrade feature is EXPERIMENTAL. While we've tested it
+extensively, we recommend backing up important files before upgrading.
+Please report any issues at: https://github.com/meizhong986/WhisperJAV/issues
+
+-------------------------------------------------------------------------------
+PREVIOUS RELEASE: v1.7.4
+-------------------------------------------------------------------------------
+Version 1.7.4 brought major improvements across 5 key areas:
+
+1. MULTI-PLATFORM GPU SUPPORT
+   - Apple Silicon Macs (M1/M2/M3/M4/M5) with native GPU acceleration
+   - NVIDIA Blackwell architecture (RTX 50-series) support
+   - Automatic device detection for best performance
+
+2. RESUME TRANSLATION FEATURE
+   - Never lose translation progress - auto-save after each batch
+   - Resume interrupted translations without re-translating
+   - Saves API costs by preserving completed work
+
+3. MULTI-LANGUAGE SUPPORT (EXPERIMENTAL)
+   - Chinese and Korean transcription
+   - Selectable source language in GUI
+   - Optimized for Japanese, basic support for others
+
+4. GUI IMPROVEMENTS
+   - Clearer language selection options
+   - CPU mode acceptance checkbox (skip GPU warnings)
+   - Fixed UI scrolling performance
+   - Enhanced drag & drop support
+
+5. BUG FIXES
+   - Windows UTF-8 encoding fix (Japanese filenames work correctly)
+   - CUDA version detection fix (accurate GPU detection)
+   - UI scrolling performance improvements
+
+-------------------------------------------------------------------------------
+WHAT THIS INSTALLER DOES
+-------------------------------------------------------------------------------
+This installer sets up a complete, self-contained WhisperJAV environment:
+
+1. Base Environment Setup (~1-2 minutes)
+   - Python 3.10.18 with conda package manager
+   - FFmpeg for audio/video processing
+   - Git for GitHub package installs
+
+2. Post-Installation Downloads (~8-15 minutes, ~2.5 GB download)
+   - PyTorch with CUDA support (for NVIDIA GPU acceleration)
+     OR CPU-only PyTorch (if no compatible GPU detected)
+   - Python dependencies (Whisper, audio processing, translation)
+   - WhisperJAV application from GitHub
+
+3. Setup and Configuration
+   - WhisperJAV-GUI.exe launcher created in installation folder
+   - Desktop shortcut creation (points to WhisperJAV-GUI.exe)
+   - Environment configuration
+   - WebView2 runtime check (required for GUI)
+
+TOTAL INSTALLATION TIME: ~10-20 minutes (depending on internet speed)
+TOTAL DISK SPACE REQUIRED: ~8 GB (install + cache + models)
+
+-------------------------------------------------------------------------------
+SYSTEM REQUIREMENTS
+-------------------------------------------------------------------------------
+MINIMUM:
+  - Windows 10 (64-bit) or Windows 11
+  - 8 GB RAM (16 GB recommended for large models)
+  - 8 GB free disk space (install + models + cache)
+  - Internet connection (for downloads)
+  - Microsoft Edge WebView2 Runtime (installer will prompt if missing)
+
+RECOMMENDED FOR BEST PERFORMANCE:
+  - NVIDIA GPU with 6+ GB VRAM (RTX 2060 or better)
+  - NVIDIA Driver with CUDA 11.8 or newer (CUDA 12.1+ preferred)
+  - 16 GB RAM
+  - SSD for faster processing
+
+SUPPORTED PLATFORMS (v1.7.5):
+  - Windows: NVIDIA GPU (RTX 20/30/40/50-series) or CPU-only
+  - macOS: Apple Silicon (M1/M2/M3/M4/M5) with native MPS acceleration
+  - macOS: Intel (CPU-only, slower)
+  - Linux: NVIDIA GPU or CPU-only
+
+CPU-ONLY MODE:
+  WhisperJAV can run without an NVIDIA GPU using CPU-only mode, but
+  processing will be significantly slower (6-10x slower than GPU mode).
+  Expect 30-60 minutes per hour of video on CPU vs 5-10 minutes on GPU.
+
+CUDA GPU ACCELERATION:
+  The installer automatically detects your NVIDIA driver's CUDA version and
+  installs the best matching PyTorch build for optimal performance:
+
+  - CUDA 11.8 - 12.0: PyTorch with CUDA 11.8 support
+  - CUDA 12.1 - 12.3: PyTorch with CUDA 12.1 support
+  - CUDA 12.4 - 12.7: PyTorch with CUDA 12.4 support
+  - CUDA 12.8+:       PyTorch with CUDA 12.4+ (Blackwell support)
+
+  This ensures you get the best possible GPU acceleration for your hardware
+  while maintaining compatibility with your driver version.
+
+-------------------------------------------------------------------------------
+AFTER INSTALLATION
+-------------------------------------------------------------------------------
+LAUNCHING THE APPLICATION:
+  - Desktop shortcut: "WhisperJAV v1.7.5.lnk" (double-click to start)
+  - Manual launch: Double-click WhisperJAV-GUI.exe in installation folder
+  - Alternative: Open the installation folder and run:
+    pythonw.exe -m whisperjav.webview_gui.main
+
+INSTALLATION LOCATION:
+  - Default: C:\Users\[YourName]\AppData\Local\WhisperJAV
+  - Or custom location chosen during installation
+
+LOGS AND DIAGNOSTICS:
+  - Installation log: install_log_v1.7.5.txt (in install folder)
+  - Application logs: Shown in GUI console during processing
+  - Failure marker: INSTALLATION_FAILED_v1.7.5.txt (only if install failed)
+
+OUTPUT FILES:
+  - Subtitles are saved next to your input video/audio files
+  - Format: SRT (SubRip Text) - compatible with most video players
+  - Filename pattern: [original_name]_output/[original_name].srt
+
+-------------------------------------------------------------------------------
+SELF-UPGRADE FEATURE (EXPERIMENTAL)
+-------------------------------------------------------------------------------
+WhisperJAV v1.7.5 introduces an experimental self-upgrade system:
+
+COMMAND LINE USAGE:
+  # Check for updates
+  whisperjav-upgrade --check
+
+  # Upgrade to latest version (with confirmation)
+  whisperjav-upgrade
+
+  # Upgrade without confirmation
+  whisperjav-upgrade --yes
+
+  # Wheel-only upgrade (skip installer dependencies)
+  whisperjav-upgrade --wheel-only
+
+GUI NOTIFICATION:
+  - The GUI will notify you when updates are available
+  - Click the notification to open the releases page
+  - Or use the CLI command to upgrade
+
+IMPORTANT NOTES:
+  - This feature is EXPERIMENTAL - backup important files first
+  - The upgrade preserves your settings and cached models
+  - If something goes wrong, reinstall from the latest installer
+  - Report issues at: https://github.com/meizhong986/WhisperJAV/issues
+
+-------------------------------------------------------------------------------
+FEATURES
+-------------------------------------------------------------------------------
+PROCESSING MODES:
+  - Balanced: Best quality, uses scene detection + voice activity detection
+  - Fast: Good quality, faster than balanced, uses scene detection
+  - Faster: Fastest processing, direct transcription without preprocessing
+
+SENSITIVITY LEVELS:
+  - Conservative: Fewer false positives, cleaner output
+  - Balanced: Good balance of detail and accuracy (recommended)
+  - Aggressive: Maximum detail capture, may include more background noise
+
+SOURCE LANGUAGES:
+  - Japanese: Fully optimized, best results
+  - Korean: Experimental, basic support
+  - Chinese: Experimental, basic support
+  - English: Experimental
+
+OUTPUT LANGUAGES:
+  - Native: Original language transcription (default)
+  - Direct to English: English transcription via Whisper
+
+ADVANCED FEATURES:
+  - Batch processing: Process multiple files sequentially
+  - Model override: Choose specific Whisper models (large-v3, large-v2, turbo)
+  - Opening credits: Add custom credit lines to subtitles
+  - Async processing: Process multiple files simultaneously (experimental)
+  - Translation: Translate subtitles to other languages (via whisperjav-translate)
+  - Resume Translation: Resume interrupted translations automatically
+  - Self-Upgrade: Update to latest version via CLI (experimental)
+
+-------------------------------------------------------------------------------
+TROUBLESHOOTING
+-------------------------------------------------------------------------------
+INSTALLATION ISSUES:
+
+1. "NVIDIA driver not found" or "CUDA version too old":
+   - Download latest driver from: https://www.nvidia.com/drivers
+   - Or accept CPU-only installation when prompted (slower processing)
+
+2. "WebView2 runtime not detected":
+   - The installer will open: https://go.microsoft.com/fwlink/p/?LinkId=2124703
+   - Download and install the "Evergreen Standalone Installer"
+   - Most Windows 10/11 systems have this pre-installed
+
+3. "Network connection failed":
+   - Check your internet connection
+   - Disable VPN or proxy temporarily
+   - Check firewall settings (allow Python, pip, git)
+
+4. "Out of disk space":
+   - Free up at least 8 GB on your system drive
+   - Models are downloaded to C:\Users\[Name]\.cache\whisper (~3 GB)
+
+5. "Installation failed after retries":
+   - Check install_log_v1.7.5.txt for specific error messages
+   - Common causes: antivirus blocking downloads, network timeout
+   - Try running the installer as Administrator (right-click > Run as admin)
+
+6. "FFmpeg not found" (FIXED in v1.7.5):
+   - This issue has been fixed in v1.7.5
+   - FFmpeg is now correctly bundled in the installer
+   - If you still see this, try reinstalling from the v1.7.5 installer
+
+RUNTIME ISSUES:
+
+1. "GUI won't launch" or "Blank window":
+   - Ensure WebView2 is installed (see issue #2 above)
+   - Try running from command line to see error messages:
+     python.exe -m whisperjav.webview_gui.main
+
+2. "Processing is very slow":
+   - Check if CUDA is enabled: Look for "GPU acceleration: ENABLED" in console
+   - If CPU-only, consider upgrading to an NVIDIA GPU
+   - For large videos, use "faster" mode for quicker results
+
+3. "Model download stuck":
+   - Large models (3GB) can take 10-20 minutes on slow connections
+   - Check your internet speed: https://fast.com
+   - Models are cached, so this only happens once
+
+4. "Subtitles have errors or gibberish":
+   - Try "balanced" mode for better quality
+   - Use "conservative" sensitivity to reduce false positives
+   - Ensure correct source language is selected
+   - For non-Japanese audio, try English or other languages
+
+5. "Application crashes during processing":
+   - Check logs in the GUI console
+   - Ensure you have enough RAM (16 GB recommended)
+   - Try processing one file at a time instead of batch
+   - Report crashes at: https://github.com/meizhong986/WhisperJAV/issues
+
+-------------------------------------------------------------------------------
+UNINSTALLING
+-------------------------------------------------------------------------------
+To completely remove WhisperJAV v1.7.5:
+
+1. Delete the installation directory:
+   C:\Users\[YourName]\AppData\Local\WhisperJAV (or your custom location)
+
+2. Delete the desktop shortcut:
+   Desktop\WhisperJAV v1.7.5.lnk
+
+3. (Optional) Delete cached models to free up ~3 GB:
+   C:\Users\[YourName]\.cache\whisper
+
+4. (Optional) Delete user configuration:
+   [Install Directory]\whisperjav_config.json
+
+NOTE: An automated uninstaller (uninstall_v1.7.5.bat) is included in the
+      installation directory for your convenience.
+
+-------------------------------------------------------------------------------
+PERFORMANCE TIPS
+-------------------------------------------------------------------------------
+1. GPU Acceleration:
+   - Ensure NVIDIA drivers are up to date
+   - Close other GPU-intensive applications before processing
+   - Verify CUDA is enabled by checking console output
+
+2. Processing Speed:
+   - "faster" mode: ~5 minutes per hour of video (GPU)
+   - "fast" mode: ~7 minutes per hour of video (GPU)
+   - "balanced" mode: ~10 minutes per hour of video (GPU)
+
+3. Quality vs Speed:
+   - For best accuracy: Use "balanced" mode with "balanced" sensitivity
+   - For speed: Use "faster" mode with "conservative" sensitivity
+   - For maximum detail: Use "balanced" mode with "aggressive" sensitivity
+
+4. Batch Processing:
+   - Process multiple files overnight using batch mode
+   - Use "async processing" (experimental) for parallel execution
+
+5. Translation Tips:
+   - Enable resume feature for long translations
+   - Progress auto-saved after each batch
+   - Re-run same command to resume interrupted translations
+
+-------------------------------------------------------------------------------
+ADVANCED USAGE
+-------------------------------------------------------------------------------
+COMMAND-LINE INTERFACE:
+  WhisperJAV also includes a CLI for automation and scripting:
+
+  # Basic transcription
+  whisperjav video.mp4 --mode balanced --sensitivity aggressive
+
+  # With translation
+  whisperjav video.mp4 --translate --target-lang english
+
+  # Batch processing
+  whisperjav folder/*.mp4 --mode fast --workers 2
+
+  Run "whisperjav --help" for full CLI documentation
+
+TRANSLATION:
+  Use whisperjav-translate for AI-powered subtitle translation:
+
+  # Translate Japanese to English
+  whisperjav-translate -i subtitles.srt --target english
+
+  # Use custom instructions
+  whisperjav-translate -i subtitles.srt --custom-gist [URL]
+
+  # Resume interrupted translation
+  whisperjav-translate -i subtitles.srt
+  # (Progress auto-saves, just re-run if interrupted)
+
+  See translation documentation for setup and API key configuration
+
+SELF-UPGRADE (EXPERIMENTAL):
+  # Check for updates
+  whisperjav-upgrade --check
+
+  # Upgrade to latest version
+  whisperjav-upgrade --yes
+
+-------------------------------------------------------------------------------
+SUPPORT & COMMUNITY
+-------------------------------------------------------------------------------
+GitHub Repository: https://github.com/meizhong986/WhisperJAV
+Issue Tracker: https://github.com/meizhong986/WhisperJAV/issues
+Documentation: See GitHub README
+
+For bugs, feature requests, or questions, please open an issue on GitHub
+with detailed information about your system and the problem.
+
+-------------------------------------------------------------------------------
+VERSION INFORMATION
+-------------------------------------------------------------------------------
+WhisperJAV Version: 1.7.5
+Release Date: January 2026
+Installer Version: v1.7.5 (conda-constructor)
+Release Type: Hotfix
+
+Key Changes in 1.7.5:
+  - Fixed missing FFmpeg issue in installer
+  - Added experimental self-upgrade feature
+  - Added update notification system
+  - Improved installer robustness
+
+Previous Major Changes (1.7.4):
+  - Multi-platform GPU support (Apple Silicon, Blackwell)
+  - Resume translation feature (auto-save progress)
+  - Multi-language support (Chinese, Korean - experimental)
+  - GUI improvements (language selection, CPU mode checkbox)
+  - Critical bug fixes (UTF-8 encoding, CUDA detection, UI scrolling)
+
+For full release notes, see RELEASE_NOTES_v1.7.5.md
+
+-------------------------------------------------------------------------------
+LICENSE
+-------------------------------------------------------------------------------
+WhisperJAV is released under the MIT License.
+See LICENSE.txt in the installation directory for details.
+
+Copyright (c) 2025-2026 MeiZhong and contributors
+
+===============================================================================

installer/generated/WhisperJAV_Launcher_v1.7.5.py

@@ -0,0 +1,165 @@
+#!/usr/bin/env python
+"""
+WhisperJAV v1.7.5 GUI Launcher
+================================
+
+This launcher script:
+1. Sets up the Python environment (PATH for ffmpeg, DLLs)
+2. Checks for first-run status (informational only)
+3. Launches the PyWebView GUI using pythonw.exe (no console window)
+4. Detaches the GUI process so the launcher can exit
+
+The GUI handles all user interaction, including model downloads
+on first transcription.
+
+No Tkinter dependency - all messages go to console (if run from terminal)
+or are handled by the GUI itself.
+"""
+
+import os
+import sys
+import json
+import subprocess
+from pathlib import Path
+
+
+def log_first_run_info(install_root: Path):
+    """
+    Check if this is first run and log informational message.
+
+    Note: We don't prompt the user here. The PyWebView GUI will handle
+    model downloads automatically when the user starts their first transcription.
+    """
+    config_path = install_root / "whisperjav_config.json"
+
+    try:
+        if config_path.exists():
+            with open(config_path, 'r', encoding='utf-8') as f:
+                config = json.load(f)
+
+            if config.get("first_run", True):
+                print("=" * 60)
+                print("  WhisperJAV First Run")
+                print("=" * 60)
+                print("AI models (~3GB) will download on first transcription.")
+                print("This is a one-time download that takes 5-10 minutes.")
+                print("=" * 60)
+                print()
+
+                # Mark first run as complete
+                config["first_run"] = False
+                with open(config_path, 'w', encoding='utf-8') as f:
+                    json.dump(config, f, indent=2)
+
+        else:
+            # Create default config for first run
+            config = {
+                "first_run": True,
+                "version": "1.7.5"
+            }
+            with open(config_path, 'w', encoding='utf-8') as f:
+                json.dump(config, f, indent=2)
+
+    except Exception as e:
+        # Non-fatal: if config handling fails, just continue
+        print(f"Note: Could not read/write config file: {e}")
+        print("This won't affect the application.")
+
+
+def main():
+    """Launch WhisperJAV GUI using pythonw.exe (windowless)"""
+    try:
+        # Determine installation root directory
+        # When frozen (PyInstaller), use directory of executable
+        # Otherwise, use sys.prefix (conda environment root)
+        if getattr(sys, 'frozen', False):
+            install_root = Path(os.path.dirname(sys.executable))
+        else:
+            install_root = Path(sys.prefix)
+
+        print(f"WhisperJAV v1.7.5 Launcher")
+        print(f"Installation: {install_root}")
+        print()
+
+        # Add Scripts and Library\bin to PATH for this session
+        # This ensures ffmpeg and other DLLs are accessible
+        scripts_dir = install_root / "Scripts"
+        lib_bin_dir = install_root / "Library" / "bin"
+
+        path_additions = [
+            str(scripts_dir),
+            str(lib_bin_dir),
+            os.environ.get("PATH", "")
+        ]
+        os.environ["PATH"] = os.pathsep.join(path_additions)
+
+        # Check for first run (informational only)
+        log_first_run_info(install_root)
+
+        # Find pythonw.exe (preferred) or python.exe (fallback)
+        pythonw = install_root / "pythonw.exe"
+        if not pythonw.exists():
+            pythonw = install_root / "python.exe"
+            if not pythonw.exists():
+                print(f"ERROR: Cannot find Python executable in {install_root}")
+                print("Installation may be corrupted.")
+                input("Press Enter to exit...")
+                sys.exit(1)
+
+        print(f"Launching GUI...")
+        print(f"Using: {pythonw}")
+        print()
+
+        # Build command to launch PyWebView GUI
+        cmd = [
+            str(pythonw),
+            "-m",
+            "whisperjav.webview_gui.main"
+        ]
+
+        # Launch as detached process (DETACHED_PROCESS flag on Windows)
+        # This allows the launcher to exit while the GUI continues running
+        if sys.platform == "win32":
+            creationflags = subprocess.DETACHED_PROCESS | subprocess.CREATE_NEW_PROCESS_GROUP
+        else:
+            creationflags = 0
+
+        subprocess.Popen(
+            cmd,
+            cwd=str(install_root),
+            creationflags=creationflags,
+            close_fds=True
+        )
+
+        print("GUI launched successfully!")
+        print("You can close this window.")
+
+    except FileNotFoundError as e:
+        print()
+        print("=" * 60)
+        print("  ERROR: File Not Found")
+        print("=" * 60)
+        print(f"{e}")
+        print()
+        print("The installation may be incomplete or corrupted.")
+        print("Please try reinstalling WhisperJAV.")
+        print("=" * 60)
+        input("Press Enter to exit...")
+        sys.exit(1)
+
+    except Exception as e:
+        print()
+        print("=" * 60)
+        print("  ERROR: Failed to Start WhisperJAV")
+        print("=" * 60)
+        print(f"{e}")
+        print()
+        print("Please check the installation or report this issue at:")
+        print("https://github.com/meizhong986/WhisperJAV/issues")
+        print("=" * 60)
+        input("Press Enter to exit...")
+        sys.exit(1)
+
+
+if __name__ == "__main__":
+    main()

installer/generated/build_installer_v1.7.5.bat

@@ -0,0 +1,289 @@
+@echo off
+REM ===============================================================================
+REM WhisperJAV v1.7.5 Installer Build Script
+REM ===============================================================================
+REM
+REM This script builds the conda-constructor installer for WhisperJAV v1.7.5
+REM
+REM Prerequisites:
+REM - Conda with constructor package installed:
+REM   conda install constructor -c conda-forge
+REM - All v1.7.5 installer files must be present in installer/ directory
+REM
+REM Build Process:
+REM 1. Check prerequisites (constructor, Python)
+REM 2. Validate installer configuration (via validate_installer_v1.7.5.py)
+REM 3. Clean previous v1.7.5 builds
+REM 4. Run constructor to build installer
+REM 5. Verify output and generate build report
+REM
+REM Output: WhisperJAV-1.7.5-Windows-x86_64.exe
+REM Logs: build_log_v1.7.5.txt
+REM
+
+SETLOCAL EnableDelayedExpansion
+
+REM ===== Configuration =====
+set CONFIG_FILE=construct_v1.7.5.yaml
+set VALIDATOR=validate_installer_v1.7.5.py
+set BUILD_LOG=build_log_v1.7.5.txt
+set VERSION=1.7.5
+
+echo ===============================================================================
+echo                    WhisperJAV v1.7.5 Installer Build
+echo ===============================================================================
+echo.
+
+REM Initialize log file
+echo WhisperJAV v1.7.5 Installer Build Log > "%BUILD_LOG%"
+echo Build started: %DATE% %TIME% >> "%BUILD_LOG%"
+echo. >> "%BUILD_LOG%"
+
+REM ===== Phase 1: Check Prerequisites =====
+echo [Phase 1/6] Checking prerequisites...
+echo [Phase 1/6] Checking prerequisites... >> "%BUILD_LOG%"
+
+REM Check if constructor is installed
+constructor --version >nul 2>nul
+if errorlevel 1 (
+    echo.
+    echo ERROR: Constructor not found or not working!
+    echo ERROR: Constructor not found >> "%BUILD_LOG%"
+    echo.
+    echo Please install constructor:
+    echo   conda install constructor -c conda-forge
+    echo.
+    echo If constructor is installed:
+    echo   1. Activate your conda environment: conda activate base
+    echo   2. Ensure conda is in your PATH
+    echo   3. Try running: constructor --version
+    echo.
+    pause
+    exit /b 1
+)
+
+for /f "tokens=*" %%a in ('constructor --version 2^>^&1') do set CONS_VER=%%a
+echo   - Constructor: %CONS_VER%
+echo   - Constructor: %CONS_VER% >> "%BUILD_LOG%"
+
+REM Check if Python is available
+python --version >nul 2>nul
+if errorlevel 1 (
+    echo.
+    echo ERROR: Python not found!
+    echo ERROR: Python not found >> "%BUILD_LOG%"
+    echo Please ensure Python is in your PATH.
+    pause
+    exit /b 1
+)
+
+for /f "tokens=*" %%a in ('python --version 2^>^&1') do set PY_VER=%%a
+echo   - Python: %PY_VER%
+echo   - Python: %PY_VER% >> "%BUILD_LOG%"
+
+REM Check if config file exists
+if not exist "%CONFIG_FILE%" (
+    echo.
+    echo ERROR: Configuration file not found: %CONFIG_FILE%
+    echo ERROR: Missing %CONFIG_FILE% >> "%BUILD_LOG%"
+    echo.
+    echo Make sure you're running this script from the installer/ directory.
+    pause
+    exit /b 1
+)
+
+echo   - Config file: %CONFIG_FILE% (found)
+echo   - Config file: %CONFIG_FILE% >> "%BUILD_LOG%"
+echo.
+
+REM ===== Phase 2: Verify WhisperJAV Wheel =====
+echo [Phase 2/6] Checking for WhisperJAV wheel...
+echo [Phase 2/6] Checking wheel... >> "%BUILD_LOG%"
+
+REM Check if wheel already exists (built by build_release.py)
+REM Use dir command for reliable wildcard detection
+dir /b whisperjav-*.whl >nul 2>nul
+if !errorlevel!==0 (
+    REM Wheel found - get the filename
+    for /f "delims=" %%f in ('dir /b whisperjav-*.whl 2^>nul') do set WHEEL_FILE=%%f
+    echo   - Found existing wheel: !WHEEL_FILE!
+    echo   - Wheel: !WHEEL_FILE! (pre-built) >> "%BUILD_LOG%"
+) else (
+    echo   - No wheel found, building from source...
+    REM Call parent directory script (uses %~dp0 so works from any location)
+    call ..\build_whisperjav_wheel.bat
+    if errorlevel 1 (
+        echo.
+        echo ERROR: Failed to build WhisperJAV wheel!
+        echo ERROR: Wheel build failed >> "%BUILD_LOG%"
+        echo.
+        pause
+        exit /b 1
+    )
+    REM Verify wheel now exists
+    dir /b whisperjav-*.whl >nul 2>nul
+    if errorlevel 1 (
+        echo.
+        echo ERROR: Wheel still not found after build!
+        echo ERROR: Wheel missing after build >> "%BUILD_LOG%"
+        echo.
+        pause
+        exit /b 1
+    )
+    for /f "delims=" %%f in ('dir /b whisperjav-*.whl 2^>nul') do set WHEEL_FILE=%%f
+    echo   - WhisperJAV wheel built: !WHEEL_FILE!
+    echo   - Wheel: !WHEEL_FILE! (local build) >> "%BUILD_LOG%"
+)
+echo.
+
+REM ===== Phase 3: Validate Configuration =====
+REM NOTE: PyInstaller build removed - WhisperJAV-GUI.exe is created by post_install
+REM       from Scripts/whisperjav-gui.exe (pip entry_points)
+echo [Phase 3/6] Validating installer configuration...
+echo [Phase 3/6] Validating configuration... >> "%BUILD_LOG%"
+
+if exist "%VALIDATOR%" (
+    python "%VALIDATOR%"
+    if errorlevel 1 (
+        echo.
+        echo ERROR: Configuration validation failed!
+        echo ERROR: Validation failed >> "%BUILD_LOG%"
+        echo Check the error messages above and fix any issues.
+        pause
+        exit /b 1
+    )
+    echo   - All validation checks passed
+    echo   - Validation: PASSED >> "%BUILD_LOG%"
+) else (
+    echo   - Validation script not found, skipping validation
+    echo   - Validation: SKIPPED (validator not found) >> "%BUILD_LOG%"
+)
+echo.
+
+REM ===== Phase 4: Clean Previous Builds =====
+echo [Phase 4/6] Cleaning previous v1.7.5 builds...
+echo [Phase 4/6] Cleaning previous builds... >> "%BUILD_LOG%"
+
+REM Only clean v1.7.5 builds, preserve other versions
+if exist build (
+    echo   - Removing build/ directory...
+    rmdir /s /q build
+    echo   - Cleaned: build/ >> "%BUILD_LOG%"
+)
+
+if exist _build (
+    echo   - Removing _build/ directory...
+    rmdir /s /q _build
+    echo   - Cleaned: _build/ >> "%BUILD_LOG%"
+)
+
+if exist WhisperJAV-%VERSION%-Windows-x86_64.exe (
+    echo   - Removing old installer: WhisperJAV-%VERSION%-Windows-x86_64.exe
+    del WhisperJAV-%VERSION%-Windows-x86_64.exe
+    echo   - Cleaned: WhisperJAV-%VERSION%-Windows-x86_64.exe >> "%BUILD_LOG%"
+)
+
+echo   - Cleanup complete
+echo.
+
+REM ===== Phase 5: Build Installer =====
+echo [Phase 5/6] Building installer with constructor...
+echo [Phase 5/6] Building installer... >> "%BUILD_LOG%"
+echo.
+echo This will take 2-5 minutes. Please wait...
+echo.
+
+REM Build with verbose output
+constructor . --config "%CONFIG_FILE%" -v
+
+if errorlevel 1 (
+    echo.
+    echo ===============================================================================
+    echo                          BUILD FAILED!
+    echo ===============================================================================
+    echo.
+    echo ERROR: Constructor build failed! >> "%BUILD_LOG%"
+    echo Check the error messages above for details.
+    echo.
+    echo Common issues:
+    echo   - Missing dependencies in construct_v1.7.5.yaml
+    echo   - Network issues downloading packages
+    echo   - Insufficient disk space
+    echo   - Corrupted conda package cache
+    echo.
+    echo To troubleshoot:
+    echo   1. Check %BUILD_LOG% for details
+    echo   2. Try: conda clean --all
+    echo   3. Verify all files in construct_v1.7.5.yaml extra_files exist
+    echo.
+    pause
+    exit /b 1
+)
+
+echo.
+
+REM ===== Phase 6: Verify Output and Generate Report =====
+echo [Phase 6/6] Verifying build output...
+echo [Phase 6/6] Verifying output... >> "%BUILD_LOG%"
+
+set INSTALLER_NAME=WhisperJAV-%VERSION%-Windows-x86_64.exe
+
+if not exist "%INSTALLER_NAME%" (
+    echo.
+    echo ERROR: Installer file was not created!
+    echo ERROR: Output file missing: %INSTALLER_NAME% >> "%BUILD_LOG%"
+    echo.
+    echo Expected: %INSTALLER_NAME%
+    echo.
+    echo This suggests the build failed silently.
+    echo Check constructor output above for errors.
+    pause
+    exit /b 1
+)
+
+REM Get file size
+for %%A in ("%INSTALLER_NAME%") do set FILESIZE=%%~zA
+set /a FILESIZE_MB=!FILESIZE! / 1048576
+
+echo   - Installer created: %INSTALLER_NAME%
+echo   - File size: !FILESIZE_MB! MB
+echo.
+
+REM Log success
+echo. >> "%BUILD_LOG%"
+echo =============================================================================== >> "%BUILD_LOG%"
+echo BUILD SUCCESSFUL >> "%BUILD_LOG%"
+echo =============================================================================== >> "%BUILD_LOG%"
+echo. >> "%BUILD_LOG%"
+echo Output file: %INSTALLER_NAME% >> "%BUILD_LOG%"
+echo File size: !FILESIZE_MB! MB >> "%BUILD_LOG%"
+echo Build completed: %DATE% %TIME% >> "%BUILD_LOG%"
+echo. >> "%BUILD_LOG%"
+
+REM ===== Build Summary =====
+echo ===============================================================================
+echo                      BUILD COMPLETED SUCCESSFULLY!
+echo ===============================================================================
+echo.
+echo Installer Details:
+echo   - File: %INSTALLER_NAME%
+echo   - Size: !FILESIZE_MB! MB
+echo   - Version: %VERSION%
+echo   - Build log: %BUILD_LOG%
+echo.
+echo Next Steps:
+echo   1. Test the installer on a clean Windows VM
+echo   2. Verify all features work correctly
+echo   3. Check installation logs (install_log_v1.7.5.txt)
+echo   4. Distribute to users
+echo.
+echo To test installer:
+echo   - Run %INSTALLER_NAME%
+echo   - Follow installation prompts
+echo   - Launch GUI from desktop shortcut
+echo   - Process a test video
+echo.
+echo ===============================================================================
+
+pause
+exit /b 0

installer/generated/build_pyinstaller/WhisperJAV.spec

@@ -0,0 +1,39 @@
+# -*- mode: python ; coding: utf-8 -*-
+
+
+a = Analysis(
+    ['..\\WhisperJAV_Launcher_v1.7.5.py'],
+    pathex=[],
+    binaries=[],
+    datas=[],
+    hiddenimports=[],
+    hookspath=[],
+    hooksconfig={},
+    runtime_hooks=[],
+    excludes=[],
+    noarchive=False,
+    optimize=0,
+)
+pyz = PYZ(a.pure)
+
+exe = EXE(
+    pyz,
+    a.scripts,
+    a.binaries,
+    a.datas,
+    [],
+    name='WhisperJAV',
+    debug=False,
+    bootloader_ignore_signals=False,
+    strip=False,
+    upx=True,
+    upx_exclude=[],
+    runtime_tmpdir=None,
+    console=False,
+    disable_windowed_traceback=False,
+    argv_emulation=False,
+    target_arch=None,
+    codesign_identity=None,
+    entitlements_file=None,
+    icon=['whisperjav_icon.ico'],
+)

installer/generated/build_pyinstaller/WhisperJAV/Analysis-00.toc

@@ -0,0 +1,902 @@
+(['C:\\BIN\\git\\WhisperJav_V1_Minami_Edition\\installer\\generated\\WhisperJAV_Launcher_v1.7.5.py'],
+ ['C:\\BIN\\git\\WhisperJav_V1_Minami_Edition\\installer\\generated'],
+ [],
+ [('C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\site-packages\\numpy\\_pyinstaller',
+   0),
+  ('C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\site-packages\\webview\\__pyinstaller',
+   0),
+  ('C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\site-packages\\_pyinstaller_hooks_contrib\\stdhooks',
+   -1000),
+  ('C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\site-packages\\_pyinstaller_hooks_contrib',
+   -1000)],
+ {},
+ [],
+ [],
+ False,
+ {},
+ 0,
+ [],
+ [],
+ '3.11.14 | packaged by conda-forge | (main, Oct 22 2025, 22:35:28) [MSC '
+ 'v.1944 64 bit (AMD64)]',
+ [('pyi_rth_inspect',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\site-packages\\PyInstaller\\hooks\\rthooks\\pyi_rth_inspect.py',
+   'PYSOURCE'),
+  ('WhisperJAV_Launcher_v1.7.5',
+   'C:\\BIN\\git\\WhisperJav_V1_Minami_Edition\\installer\\generated\\WhisperJAV_Launcher_v1.7.5.py',
+   'PYSOURCE')],
+ [('zipfile',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\zipfile.py',
+   'PYMODULE'),
+  ('argparse',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\argparse.py',
+   'PYMODULE'),
+  ('textwrap',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\textwrap.py',
+   'PYMODULE'),
+  ('copy', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\copy.py', 'PYMODULE'),
+  ('gettext',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\gettext.py',
+   'PYMODULE'),
+  ('py_compile',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\py_compile.py',
+   'PYMODULE'),
+  ('importlib.machinery',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\machinery.py',
+   'PYMODULE'),
+  ('importlib',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\__init__.py',
+   'PYMODULE'),
+  ('importlib._bootstrap',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\_bootstrap.py',
+   'PYMODULE'),
+  ('importlib._bootstrap_external',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\_bootstrap_external.py',
+   'PYMODULE'),
+  ('importlib.metadata',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\metadata\\__init__.py',
+   'PYMODULE'),
+  ('typing', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\typing.py', 'PYMODULE'),
+  ('importlib.abc',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\abc.py',
+   'PYMODULE'),
+  ('importlib.resources.abc',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\resources\\abc.py',
+   'PYMODULE'),
+  ('importlib.resources',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\resources\\__init__.py',
+   'PYMODULE'),
+  ('importlib.resources._legacy',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\resources\\_legacy.py',
+   'PYMODULE'),
+  ('importlib.resources._common',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\resources\\_common.py',
+   'PYMODULE'),
+  ('importlib.resources._adapters',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\resources\\_adapters.py',
+   'PYMODULE'),
+  ('tempfile',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\tempfile.py',
+   'PYMODULE'),
+  ('random', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\random.py', 'PYMODULE'),
+  ('statistics',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\statistics.py',
+   'PYMODULE'),
+  ('decimal',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\decimal.py',
+   'PYMODULE'),
+  ('_pydecimal',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\_pydecimal.py',
+   'PYMODULE'),
+  ('contextvars',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\contextvars.py',
+   'PYMODULE'),
+  ('fractions',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\fractions.py',
+   'PYMODULE'),
+  ('numbers',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\numbers.py',
+   'PYMODULE'),
+  ('hashlib',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\hashlib.py',
+   'PYMODULE'),
+  ('logging',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\logging\\__init__.py',
+   'PYMODULE'),
+  ('pickle', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\pickle.py', 'PYMODULE'),
+  ('pprint', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\pprint.py', 'PYMODULE'),
+  ('dataclasses',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\dataclasses.py',
+   'PYMODULE'),
+  ('_compat_pickle',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\_compat_pickle.py',
+   'PYMODULE'),
+  ('string', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\string.py', 'PYMODULE'),
+  ('bisect', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\bisect.py', 'PYMODULE'),
+  ('importlib._abc',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\_abc.py',
+   'PYMODULE'),
+  ('importlib.metadata._itertools',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\metadata\\_itertools.py',
+   'PYMODULE'),
+  ('importlib.metadata._functools',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\metadata\\_functools.py',
+   'PYMODULE'),
+  ('importlib.metadata._collections',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\metadata\\_collections.py',
+   'PYMODULE'),
+  ('importlib.metadata._meta',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\metadata\\_meta.py',
+   'PYMODULE'),
+  ('importlib.metadata._adapters',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\metadata\\_adapters.py',
+   'PYMODULE'),
+  ('importlib.metadata._text',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\metadata\\_text.py',
+   'PYMODULE'),
+  ('email.message',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\message.py',
+   'PYMODULE'),
+  ('email.policy',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\policy.py',
+   'PYMODULE'),
+  ('email.contentmanager',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\contentmanager.py',
+   'PYMODULE'),
+  ('email.quoprimime',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\quoprimime.py',
+   'PYMODULE'),
+  ('email.headerregistry',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\headerregistry.py',
+   'PYMODULE'),
+  ('email._header_value_parser',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\_header_value_parser.py',
+   'PYMODULE'),
+  ('urllib',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\urllib\\__init__.py',
+   'PYMODULE'),
+  ('email.iterators',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\iterators.py',
+   'PYMODULE'),
+  ('email.generator',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\generator.py',
+   'PYMODULE'),
+  ('email._encoded_words',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\_encoded_words.py',
+   'PYMODULE'),
+  ('base64', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\base64.py', 'PYMODULE'),
+  ('getopt', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\getopt.py', 'PYMODULE'),
+  ('email.charset',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\charset.py',
+   'PYMODULE'),
+  ('email.encoders',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\encoders.py',
+   'PYMODULE'),
+  ('email.base64mime',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\base64mime.py',
+   'PYMODULE'),
+  ('email._policybase',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\_policybase.py',
+   'PYMODULE'),
+  ('email.header',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\header.py',
+   'PYMODULE'),
+  ('email.errors',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\errors.py',
+   'PYMODULE'),
+  ('email.utils',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\utils.py',
+   'PYMODULE'),
+  ('email._parseaddr',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\_parseaddr.py',
+   'PYMODULE'),
+  ('calendar',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\calendar.py',
+   'PYMODULE'),
+  ('urllib.parse',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\urllib\\parse.py',
+   'PYMODULE'),
+  ('ipaddress',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\ipaddress.py',
+   'PYMODULE'),
+  ('datetime',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\datetime.py',
+   'PYMODULE'),
+  ('_strptime',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\_strptime.py',
+   'PYMODULE'),
+  ('socket', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\socket.py', 'PYMODULE'),
+  ('selectors',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\selectors.py',
+   'PYMODULE'),
+  ('quopri', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\quopri.py', 'PYMODULE'),
+  ('email',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\__init__.py',
+   'PYMODULE'),
+  ('email.parser',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\parser.py',
+   'PYMODULE'),
+  ('email.feedparser',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\feedparser.py',
+   'PYMODULE'),
+  ('csv', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\csv.py', 'PYMODULE'),
+  ('importlib.readers',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\readers.py',
+   'PYMODULE'),
+  ('importlib.resources.readers',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\resources\\readers.py',
+   'PYMODULE'),
+  ('importlib.resources._itertools',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\resources\\_itertools.py',
+   'PYMODULE'),
+  ('tokenize',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\tokenize.py',
+   'PYMODULE'),
+  ('token', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\token.py', 'PYMODULE'),
+  ('lzma', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\lzma.py', 'PYMODULE'),
+  ('_compression',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\_compression.py',
+   'PYMODULE'),
+  ('bz2', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\bz2.py', 'PYMODULE'),
+  ('contextlib',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\contextlib.py',
+   'PYMODULE'),
+  ('threading',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\threading.py',
+   'PYMODULE'),
+  ('_threading_local',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\_threading_local.py',
+   'PYMODULE'),
+  ('struct', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\struct.py', 'PYMODULE'),
+  ('shutil', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\shutil.py', 'PYMODULE'),
+  ('tarfile',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\tarfile.py',
+   'PYMODULE'),
+  ('gzip', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\gzip.py', 'PYMODULE'),
+  ('fnmatch',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\fnmatch.py',
+   'PYMODULE'),
+  ('importlib.util',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\util.py',
+   'PYMODULE'),
+  ('inspect',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\inspect.py',
+   'PYMODULE'),
+  ('dis', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\dis.py', 'PYMODULE'),
+  ('opcode', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\opcode.py', 'PYMODULE'),
+  ('ast', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\ast.py', 'PYMODULE'),
+  ('_py_abc',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\_py_abc.py',
+   'PYMODULE'),
+  ('tracemalloc',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\tracemalloc.py',
+   'PYMODULE'),
+  ('stringprep',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\stringprep.py',
+   'PYMODULE'),
+  ('pathlib',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\pathlib.py',
+   'PYMODULE'),
+  ('subprocess',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\subprocess.py',
+   'PYMODULE'),
+  ('signal', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\signal.py', 'PYMODULE'),
+  ('json',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\json\\__init__.py',
+   'PYMODULE'),
+  ('json.encoder',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\json\\encoder.py',
+   'PYMODULE'),
+  ('json.decoder',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\json\\decoder.py',
+   'PYMODULE'),
+  ('json.scanner',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\json\\scanner.py',
+   'PYMODULE')],
+ [('python311.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\python311.dll',
+   'BINARY'),
+  ('_decimal.pyd',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\DLLs\\_decimal.pyd',
+   'EXTENSION'),
+  ('_hashlib.pyd',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\DLLs\\_hashlib.pyd',
+   'EXTENSION'),
+  ('unicodedata.pyd',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\DLLs\\unicodedata.pyd',
+   'EXTENSION'),
+  ('select.pyd',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\DLLs\\select.pyd',
+   'EXTENSION'),
+  ('_socket.pyd',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\DLLs\\_socket.pyd',
+   'EXTENSION'),
+  ('_lzma.pyd',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\DLLs\\_lzma.pyd',
+   'EXTENSION'),
+  ('_bz2.pyd',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\DLLs\\_bz2.pyd',
+   'EXTENSION'),
+  ('api-ms-win-crt-runtime-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-crt-runtime-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-crt-heap-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-crt-heap-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-crt-stdio-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-crt-stdio-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-crt-conio-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-crt-conio-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-crt-math-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-crt-math-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-crt-environment-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-crt-environment-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-crt-time-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-crt-time-l1-1-0.dll',
+   'BINARY'),
+  ('zlib.dll', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\zlib.dll', 'BINARY'),
+  ('api-ms-win-crt-string-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-crt-string-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-crt-process-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-crt-process-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-crt-filesystem-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-crt-filesystem-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-crt-locale-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-crt-locale-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-crt-convert-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-crt-convert-l1-1-0.dll',
+   'BINARY'),
+  ('VCRUNTIME140.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\VCRUNTIME140.dll',
+   'BINARY'),
+  ('libcrypto-3-x64.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Library\\bin\\libcrypto-3-x64.dll',
+   'BINARY'),
+  ('liblzma.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Library\\bin\\liblzma.dll',
+   'BINARY'),
+  ('libbz2.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Library\\bin\\libbz2.dll',
+   'BINARY'),
+  ('ucrtbase.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\ucrtbase.dll',
+   'BINARY'),
+  ('api-ms-win-crt-utility-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Library\\bin\\api-ms-win-crt-utility-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-sysinfo-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-sysinfo-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-errorhandling-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-errorhandling-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-console-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-console-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-localization-l1-2-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-localization-l1-2-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-memory-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-memory-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-file-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-file-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-profile-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-profile-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-synch-l1-2-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-synch-l1-2-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-util-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-util-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-string-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-string-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-file-l1-2-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-file-l1-2-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-processenvironment-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-processenvironment-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-timezone-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-timezone-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-synch-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-synch-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-handle-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-handle-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-libraryloader-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-libraryloader-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-datetime-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-datetime-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-interlocked-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-interlocked-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-heap-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-heap-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-namedpipe-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-namedpipe-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-fibers-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-fibers-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-processthreads-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-processthreads-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-processthreads-l1-1-1.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-processthreads-l1-1-1.dll',
+   'BINARY'),
+  ('api-ms-win-core-debug-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-debug-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-file-l2-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-file-l2-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-rtlsupport-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-rtlsupport-l1-1-0.dll',
+   'BINARY')],
+ [],
+ [],
+ [('base_library.zip',
+   'C:\\BIN\\git\\WhisperJav_V1_Minami_Edition\\installer\\generated\\build_pyinstaller\\WhisperJAV\\base_library.zip',
+   'DATA')],
+ [('ntpath', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\ntpath.py', 'PYMODULE'),
+  ('types', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\types.py', 'PYMODULE'),
+  ('keyword',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\keyword.py',
+   'PYMODULE'),
+  ('copyreg',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\copyreg.py',
+   'PYMODULE'),
+  ('sre_constants',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\sre_constants.py',
+   'PYMODULE'),
+  ('operator',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\operator.py',
+   'PYMODULE'),
+  ('genericpath',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\genericpath.py',
+   'PYMODULE'),
+  ('locale', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\locale.py', 'PYMODULE'),
+  ('abc', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\abc.py', 'PYMODULE'),
+  ('collections.abc',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\collections\\abc.py',
+   'PYMODULE'),
+  ('collections',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\collections\\__init__.py',
+   'PYMODULE'),
+  ('enum', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\enum.py', 'PYMODULE'),
+  ('functools',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\functools.py',
+   'PYMODULE'),
+  ('reprlib',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\reprlib.py',
+   'PYMODULE'),
+  ('warnings',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\warnings.py',
+   'PYMODULE'),
+  ('encodings.zlib_codec',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\zlib_codec.py',
+   'PYMODULE'),
+  ('encodings.uu_codec',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\uu_codec.py',
+   'PYMODULE'),
+  ('encodings.utf_8_sig',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\utf_8_sig.py',
+   'PYMODULE'),
+  ('encodings.utf_8',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\utf_8.py',
+   'PYMODULE'),
+  ('encodings.utf_7',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\utf_7.py',
+   'PYMODULE'),
+  ('encodings.utf_32_le',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\utf_32_le.py',
+   'PYMODULE'),
+  ('encodings.utf_32_be',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\utf_32_be.py',
+   'PYMODULE'),
+  ('encodings.utf_32',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\utf_32.py',
+   'PYMODULE'),
+  ('encodings.utf_16_le',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\utf_16_le.py',
+   'PYMODULE'),
+  ('encodings.utf_16_be',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\utf_16_be.py',
+   'PYMODULE'),
+  ('encodings.utf_16',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\utf_16.py',
+   'PYMODULE'),
+  ('encodings.unicode_escape',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\unicode_escape.py',
+   'PYMODULE'),
+  ('encodings.undefined',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\undefined.py',
+   'PYMODULE'),
+  ('encodings.tis_620',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\tis_620.py',
+   'PYMODULE'),
+  ('encodings.shift_jisx0213',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\shift_jisx0213.py',
+   'PYMODULE'),
+  ('encodings.shift_jis_2004',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\shift_jis_2004.py',
+   'PYMODULE'),
+  ('encodings.shift_jis',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\shift_jis.py',
+   'PYMODULE'),
+  ('encodings.rot_13',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\rot_13.py',
+   'PYMODULE'),
+  ('encodings.raw_unicode_escape',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\raw_unicode_escape.py',
+   'PYMODULE'),
+  ('encodings.quopri_codec',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\quopri_codec.py',
+   'PYMODULE'),
+  ('encodings.punycode',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\punycode.py',
+   'PYMODULE'),
+  ('encodings.ptcp154',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\ptcp154.py',
+   'PYMODULE'),
+  ('encodings.palmos',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\palmos.py',
+   'PYMODULE'),
+  ('encodings.oem',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\oem.py',
+   'PYMODULE'),
+  ('encodings.mbcs',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\mbcs.py',
+   'PYMODULE'),
+  ('encodings.mac_turkish',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\mac_turkish.py',
+   'PYMODULE'),
+  ('encodings.mac_romanian',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\mac_romanian.py',
+   'PYMODULE'),
+  ('encodings.mac_roman',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\mac_roman.py',
+   'PYMODULE'),
+  ('encodings.mac_latin2',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\mac_latin2.py',
+   'PYMODULE'),
+  ('encodings.mac_iceland',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\mac_iceland.py',
+   'PYMODULE'),
+  ('encodings.mac_greek',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\mac_greek.py',
+   'PYMODULE'),
+  ('encodings.mac_farsi',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\mac_farsi.py',
+   'PYMODULE'),
+  ('encodings.mac_cyrillic',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\mac_cyrillic.py',
+   'PYMODULE'),
+  ('encodings.mac_croatian',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\mac_croatian.py',
+   'PYMODULE'),
+  ('encodings.mac_arabic',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\mac_arabic.py',
+   'PYMODULE'),
+  ('encodings.latin_1',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\latin_1.py',
+   'PYMODULE'),
+  ('encodings.kz1048',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\kz1048.py',
+   'PYMODULE'),
+  ('encodings.koi8_u',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\koi8_u.py',
+   'PYMODULE'),
+  ('encodings.koi8_t',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\koi8_t.py',
+   'PYMODULE'),
+  ('encodings.koi8_r',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\koi8_r.py',
+   'PYMODULE'),
+  ('encodings.johab',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\johab.py',
+   'PYMODULE'),
+  ('encodings.iso8859_9',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\iso8859_9.py',
+   'PYMODULE'),
+  ('encodings.iso8859_8',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\iso8859_8.py',
+   'PYMODULE'),
+  ('encodings.iso8859_7',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\iso8859_7.py',
+   'PYMODULE'),
+  ('encodings.iso8859_6',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\iso8859_6.py',
+   'PYMODULE'),
+  ('encodings.iso8859_5',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\iso8859_5.py',
+   'PYMODULE'),
+  ('encodings.iso8859_4',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\iso8859_4.py',
+   'PYMODULE'),
+  ('encodings.iso8859_3',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\iso8859_3.py',
+   'PYMODULE'),
+  ('encodings.iso8859_2',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\iso8859_2.py',
+   'PYMODULE'),
+  ('encodings.iso8859_16',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\iso8859_16.py',
+   'PYMODULE'),
+  ('encodings.iso8859_15',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\iso8859_15.py',
+   'PYMODULE'),
+  ('encodings.iso8859_14',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\iso8859_14.py',
+   'PYMODULE'),
+  ('encodings.iso8859_13',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\iso8859_13.py',
+   'PYMODULE'),
+  ('encodings.iso8859_11',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\iso8859_11.py',
+   'PYMODULE'),
+  ('encodings.iso8859_10',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\iso8859_10.py',
+   'PYMODULE'),
+  ('encodings.iso8859_1',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\iso8859_1.py',
+   'PYMODULE'),
+  ('encodings.iso2022_kr',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\iso2022_kr.py',
+   'PYMODULE'),
+  ('encodings.iso2022_jp_ext',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\iso2022_jp_ext.py',
+   'PYMODULE'),
+  ('encodings.iso2022_jp_3',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\iso2022_jp_3.py',
+   'PYMODULE'),
+  ('encodings.iso2022_jp_2004',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\iso2022_jp_2004.py',
+   'PYMODULE'),
+  ('encodings.iso2022_jp_2',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\iso2022_jp_2.py',
+   'PYMODULE'),
+  ('encodings.iso2022_jp_1',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\iso2022_jp_1.py',
+   'PYMODULE'),
+  ('encodings.iso2022_jp',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\iso2022_jp.py',
+   'PYMODULE'),
+  ('encodings.idna',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\idna.py',
+   'PYMODULE'),
+  ('encodings.hz',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\hz.py',
+   'PYMODULE'),
+  ('encodings.hp_roman8',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\hp_roman8.py',
+   'PYMODULE'),
+  ('encodings.hex_codec',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\hex_codec.py',
+   'PYMODULE'),
+  ('encodings.gbk',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\gbk.py',
+   'PYMODULE'),
+  ('encodings.gb2312',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\gb2312.py',
+   'PYMODULE'),
+  ('encodings.gb18030',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\gb18030.py',
+   'PYMODULE'),
+  ('encodings.euc_kr',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\euc_kr.py',
+   'PYMODULE'),
+  ('encodings.euc_jp',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\euc_jp.py',
+   'PYMODULE'),
+  ('encodings.euc_jisx0213',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\euc_jisx0213.py',
+   'PYMODULE'),
+  ('encodings.euc_jis_2004',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\euc_jis_2004.py',
+   'PYMODULE'),
+  ('encodings.cp950',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp950.py',
+   'PYMODULE'),
+  ('encodings.cp949',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp949.py',
+   'PYMODULE'),
+  ('encodings.cp932',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp932.py',
+   'PYMODULE'),
+  ('encodings.cp875',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp875.py',
+   'PYMODULE'),
+  ('encodings.cp874',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp874.py',
+   'PYMODULE'),
+  ('encodings.cp869',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp869.py',
+   'PYMODULE'),
+  ('encodings.cp866',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp866.py',
+   'PYMODULE'),
+  ('encodings.cp865',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp865.py',
+   'PYMODULE'),
+  ('encodings.cp864',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp864.py',
+   'PYMODULE'),
+  ('encodings.cp863',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp863.py',
+   'PYMODULE'),
+  ('encodings.cp862',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp862.py',
+   'PYMODULE'),
+  ('encodings.cp861',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp861.py',
+   'PYMODULE'),
+  ('encodings.cp860',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp860.py',
+   'PYMODULE'),
+  ('encodings.cp858',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp858.py',
+   'PYMODULE'),
+  ('encodings.cp857',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp857.py',
+   'PYMODULE'),
+  ('encodings.cp856',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp856.py',
+   'PYMODULE'),
+  ('encodings.cp855',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp855.py',
+   'PYMODULE'),
+  ('encodings.cp852',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp852.py',
+   'PYMODULE'),
+  ('encodings.cp850',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp850.py',
+   'PYMODULE'),
+  ('encodings.cp775',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp775.py',
+   'PYMODULE'),
+  ('encodings.cp737',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp737.py',
+   'PYMODULE'),
+  ('encodings.cp720',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp720.py',
+   'PYMODULE'),
+  ('encodings.cp500',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp500.py',
+   'PYMODULE'),
+  ('encodings.cp437',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp437.py',
+   'PYMODULE'),
+  ('encodings.cp424',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp424.py',
+   'PYMODULE'),
+  ('encodings.cp273',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp273.py',
+   'PYMODULE'),
+  ('encodings.cp1258',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp1258.py',
+   'PYMODULE'),
+  ('encodings.cp1257',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp1257.py',
+   'PYMODULE'),
+  ('encodings.cp1256',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp1256.py',
+   'PYMODULE'),
+  ('encodings.cp1255',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp1255.py',
+   'PYMODULE'),
+  ('encodings.cp1254',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp1254.py',
+   'PYMODULE'),
+  ('encodings.cp1253',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp1253.py',
+   'PYMODULE'),
+  ('encodings.cp1252',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp1252.py',
+   'PYMODULE'),
+  ('encodings.cp1251',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp1251.py',
+   'PYMODULE'),
+  ('encodings.cp1250',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp1250.py',
+   'PYMODULE'),
+  ('encodings.cp1140',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp1140.py',
+   'PYMODULE'),
+  ('encodings.cp1125',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp1125.py',
+   'PYMODULE'),
+  ('encodings.cp1026',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp1026.py',
+   'PYMODULE'),
+  ('encodings.cp1006',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp1006.py',
+   'PYMODULE'),
+  ('encodings.cp037',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\cp037.py',
+   'PYMODULE'),
+  ('encodings.charmap',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\charmap.py',
+   'PYMODULE'),
+  ('encodings.bz2_codec',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\bz2_codec.py',
+   'PYMODULE'),
+  ('encodings.big5hkscs',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\big5hkscs.py',
+   'PYMODULE'),
+  ('encodings.big5',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\big5.py',
+   'PYMODULE'),
+  ('encodings.base64_codec',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\base64_codec.py',
+   'PYMODULE'),
+  ('encodings.ascii',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\ascii.py',
+   'PYMODULE'),
+  ('encodings.aliases',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\aliases.py',
+   'PYMODULE'),
+  ('encodings',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\encodings\\__init__.py',
+   'PYMODULE'),
+  ('posixpath',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\posixpath.py',
+   'PYMODULE'),
+  ('traceback',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\traceback.py',
+   'PYMODULE'),
+  ('heapq', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\heapq.py', 'PYMODULE'),
+  ('re._parser',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\re\\_parser.py',
+   'PYMODULE'),
+  ('re._constants',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\re\\_constants.py',
+   'PYMODULE'),
+  ('re._compiler',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\re\\_compiler.py',
+   'PYMODULE'),
+  ('re._casefix',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\re\\_casefix.py',
+   'PYMODULE'),
+  ('re',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\re\\__init__.py',
+   'PYMODULE'),
+  ('io', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\io.py', 'PYMODULE'),
+  ('_weakrefset',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\_weakrefset.py',
+   'PYMODULE'),
+  ('codecs', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\codecs.py', 'PYMODULE'),
+  ('linecache',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\linecache.py',
+   'PYMODULE'),
+  ('weakref',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\weakref.py',
+   'PYMODULE'),
+  ('_collections_abc',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\_collections_abc.py',
+   'PYMODULE'),
+  ('stat', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\stat.py', 'PYMODULE'),
+  ('sre_compile',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\sre_compile.py',
+   'PYMODULE'),
+  ('sre_parse',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\sre_parse.py',
+   'PYMODULE'),
+  ('os', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\os.py', 'PYMODULE')])

installer/generated/build_pyinstaller/WhisperJAV/PKG-00.toc

@@ -0,0 +1,206 @@
+('C:\\BIN\\git\\WhisperJav_V1_Minami_Edition\\installer\\generated\\build_pyinstaller\\WhisperJAV\\WhisperJAV.pkg',
+ {'BINARY': True,
+  'DATA': True,
+  'EXECUTABLE': True,
+  'EXTENSION': True,
+  'PYMODULE': True,
+  'PYSOURCE': True,
+  'PYZ': False,
+  'SPLASH': True,
+  'SYMLINK': False},
+ [('pyi-contents-directory _internal', '', 'OPTION'),
+  ('PYZ-00.pyz',
+   'C:\\BIN\\git\\WhisperJav_V1_Minami_Edition\\installer\\generated\\build_pyinstaller\\WhisperJAV\\PYZ-00.pyz',
+   'PYZ'),
+  ('struct',
+   'C:\\BIN\\git\\WhisperJav_V1_Minami_Edition\\installer\\generated\\build_pyinstaller\\WhisperJAV\\localpycs\\struct.pyc',
+   'PYMODULE'),
+  ('pyimod01_archive',
+   'C:\\BIN\\git\\WhisperJav_V1_Minami_Edition\\installer\\generated\\build_pyinstaller\\WhisperJAV\\localpycs\\pyimod01_archive.pyc',
+   'PYMODULE'),
+  ('pyimod02_importers',
+   'C:\\BIN\\git\\WhisperJav_V1_Minami_Edition\\installer\\generated\\build_pyinstaller\\WhisperJAV\\localpycs\\pyimod02_importers.pyc',
+   'PYMODULE'),
+  ('pyimod03_ctypes',
+   'C:\\BIN\\git\\WhisperJav_V1_Minami_Edition\\installer\\generated\\build_pyinstaller\\WhisperJAV\\localpycs\\pyimod03_ctypes.pyc',
+   'PYMODULE'),
+  ('pyimod04_pywin32',
+   'C:\\BIN\\git\\WhisperJav_V1_Minami_Edition\\installer\\generated\\build_pyinstaller\\WhisperJAV\\localpycs\\pyimod04_pywin32.pyc',
+   'PYMODULE'),
+  ('pyiboot01_bootstrap',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\site-packages\\PyInstaller\\loader\\pyiboot01_bootstrap.py',
+   'PYSOURCE'),
+  ('pyi_rth_inspect',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\site-packages\\PyInstaller\\hooks\\rthooks\\pyi_rth_inspect.py',
+   'PYSOURCE'),
+  ('WhisperJAV_Launcher_v1.7.5',
+   'C:\\BIN\\git\\WhisperJav_V1_Minami_Edition\\installer\\generated\\WhisperJAV_Launcher_v1.7.5.py',
+   'PYSOURCE'),
+  ('python311.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\python311.dll',
+   'BINARY'),
+  ('_decimal.pyd',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\DLLs\\_decimal.pyd',
+   'EXTENSION'),
+  ('_hashlib.pyd',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\DLLs\\_hashlib.pyd',
+   'EXTENSION'),
+  ('unicodedata.pyd',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\DLLs\\unicodedata.pyd',
+   'EXTENSION'),
+  ('select.pyd',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\DLLs\\select.pyd',
+   'EXTENSION'),
+  ('_socket.pyd',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\DLLs\\_socket.pyd',
+   'EXTENSION'),
+  ('_lzma.pyd',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\DLLs\\_lzma.pyd',
+   'EXTENSION'),
+  ('_bz2.pyd',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\DLLs\\_bz2.pyd',
+   'EXTENSION'),
+  ('api-ms-win-crt-runtime-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-crt-runtime-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-crt-heap-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-crt-heap-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-crt-stdio-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-crt-stdio-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-crt-conio-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-crt-conio-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-crt-math-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-crt-math-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-crt-environment-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-crt-environment-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-crt-time-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-crt-time-l1-1-0.dll',
+   'BINARY'),
+  ('zlib.dll', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\zlib.dll', 'BINARY'),
+  ('api-ms-win-crt-string-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-crt-string-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-crt-process-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-crt-process-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-crt-filesystem-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-crt-filesystem-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-crt-locale-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-crt-locale-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-crt-convert-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-crt-convert-l1-1-0.dll',
+   'BINARY'),
+  ('VCRUNTIME140.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\VCRUNTIME140.dll',
+   'BINARY'),
+  ('libcrypto-3-x64.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Library\\bin\\libcrypto-3-x64.dll',
+   'BINARY'),
+  ('liblzma.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Library\\bin\\liblzma.dll',
+   'BINARY'),
+  ('libbz2.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Library\\bin\\libbz2.dll',
+   'BINARY'),
+  ('ucrtbase.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\ucrtbase.dll',
+   'BINARY'),
+  ('api-ms-win-crt-utility-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Library\\bin\\api-ms-win-crt-utility-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-sysinfo-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-sysinfo-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-errorhandling-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-errorhandling-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-console-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-console-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-localization-l1-2-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-localization-l1-2-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-memory-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-memory-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-file-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-file-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-profile-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-profile-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-synch-l1-2-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-synch-l1-2-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-util-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-util-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-string-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-string-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-file-l1-2-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-file-l1-2-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-processenvironment-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-processenvironment-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-timezone-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-timezone-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-synch-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-synch-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-handle-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-handle-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-libraryloader-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-libraryloader-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-datetime-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-datetime-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-interlocked-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-interlocked-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-heap-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-heap-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-namedpipe-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-namedpipe-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-fibers-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-fibers-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-processthreads-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-processthreads-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-processthreads-l1-1-1.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-processthreads-l1-1-1.dll',
+   'BINARY'),
+  ('api-ms-win-core-debug-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-debug-l1-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-file-l2-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-file-l2-1-0.dll',
+   'BINARY'),
+  ('api-ms-win-core-rtlsupport-l1-1-0.dll',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\api-ms-win-core-rtlsupport-l1-1-0.dll',
+   'BINARY'),
+  ('base_library.zip',
+   'C:\\BIN\\git\\WhisperJav_V1_Minami_Edition\\installer\\generated\\build_pyinstaller\\WhisperJAV\\base_library.zip',
+   'DATA')],
+ 'python311.dll',
+ False,
+ False,
+ False,
+ [],
+ None,
+ None,
+ None)

installer/generated/build_pyinstaller/WhisperJAV/PYZ-00.toc

@@ -0,0 +1,268 @@
+('C:\\BIN\\git\\WhisperJav_V1_Minami_Edition\\installer\\generated\\build_pyinstaller\\WhisperJAV\\PYZ-00.pyz',
+ [('_compat_pickle',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\_compat_pickle.py',
+   'PYMODULE'),
+  ('_compression',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\_compression.py',
+   'PYMODULE'),
+  ('_py_abc',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\_py_abc.py',
+   'PYMODULE'),
+  ('_pydecimal',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\_pydecimal.py',
+   'PYMODULE'),
+  ('_strptime',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\_strptime.py',
+   'PYMODULE'),
+  ('_threading_local',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\_threading_local.py',
+   'PYMODULE'),
+  ('argparse',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\argparse.py',
+   'PYMODULE'),
+  ('ast', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\ast.py', 'PYMODULE'),
+  ('base64', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\base64.py', 'PYMODULE'),
+  ('bisect', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\bisect.py', 'PYMODULE'),
+  ('bz2', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\bz2.py', 'PYMODULE'),
+  ('calendar',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\calendar.py',
+   'PYMODULE'),
+  ('contextlib',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\contextlib.py',
+   'PYMODULE'),
+  ('contextvars',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\contextvars.py',
+   'PYMODULE'),
+  ('copy', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\copy.py', 'PYMODULE'),
+  ('csv', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\csv.py', 'PYMODULE'),
+  ('dataclasses',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\dataclasses.py',
+   'PYMODULE'),
+  ('datetime',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\datetime.py',
+   'PYMODULE'),
+  ('decimal',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\decimal.py',
+   'PYMODULE'),
+  ('dis', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\dis.py', 'PYMODULE'),
+  ('email',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\__init__.py',
+   'PYMODULE'),
+  ('email._encoded_words',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\_encoded_words.py',
+   'PYMODULE'),
+  ('email._header_value_parser',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\_header_value_parser.py',
+   'PYMODULE'),
+  ('email._parseaddr',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\_parseaddr.py',
+   'PYMODULE'),
+  ('email._policybase',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\_policybase.py',
+   'PYMODULE'),
+  ('email.base64mime',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\base64mime.py',
+   'PYMODULE'),
+  ('email.charset',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\charset.py',
+   'PYMODULE'),
+  ('email.contentmanager',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\contentmanager.py',
+   'PYMODULE'),
+  ('email.encoders',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\encoders.py',
+   'PYMODULE'),
+  ('email.errors',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\errors.py',
+   'PYMODULE'),
+  ('email.feedparser',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\feedparser.py',
+   'PYMODULE'),
+  ('email.generator',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\generator.py',
+   'PYMODULE'),
+  ('email.header',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\header.py',
+   'PYMODULE'),
+  ('email.headerregistry',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\headerregistry.py',
+   'PYMODULE'),
+  ('email.iterators',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\iterators.py',
+   'PYMODULE'),
+  ('email.message',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\message.py',
+   'PYMODULE'),
+  ('email.parser',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\parser.py',
+   'PYMODULE'),
+  ('email.policy',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\policy.py',
+   'PYMODULE'),
+  ('email.quoprimime',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\quoprimime.py',
+   'PYMODULE'),
+  ('email.utils',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\email\\utils.py',
+   'PYMODULE'),
+  ('fnmatch',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\fnmatch.py',
+   'PYMODULE'),
+  ('fractions',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\fractions.py',
+   'PYMODULE'),
+  ('getopt', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\getopt.py', 'PYMODULE'),
+  ('gettext',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\gettext.py',
+   'PYMODULE'),
+  ('gzip', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\gzip.py', 'PYMODULE'),
+  ('hashlib',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\hashlib.py',
+   'PYMODULE'),
+  ('importlib',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\__init__.py',
+   'PYMODULE'),
+  ('importlib._abc',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\_abc.py',
+   'PYMODULE'),
+  ('importlib._bootstrap',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\_bootstrap.py',
+   'PYMODULE'),
+  ('importlib._bootstrap_external',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\_bootstrap_external.py',
+   'PYMODULE'),
+  ('importlib.abc',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\abc.py',
+   'PYMODULE'),
+  ('importlib.machinery',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\machinery.py',
+   'PYMODULE'),
+  ('importlib.metadata',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\metadata\\__init__.py',
+   'PYMODULE'),
+  ('importlib.metadata._adapters',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\metadata\\_adapters.py',
+   'PYMODULE'),
+  ('importlib.metadata._collections',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\metadata\\_collections.py',
+   'PYMODULE'),
+  ('importlib.metadata._functools',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\metadata\\_functools.py',
+   'PYMODULE'),
+  ('importlib.metadata._itertools',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\metadata\\_itertools.py',
+   'PYMODULE'),
+  ('importlib.metadata._meta',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\metadata\\_meta.py',
+   'PYMODULE'),
+  ('importlib.metadata._text',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\metadata\\_text.py',
+   'PYMODULE'),
+  ('importlib.readers',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\readers.py',
+   'PYMODULE'),
+  ('importlib.resources',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\resources\\__init__.py',
+   'PYMODULE'),
+  ('importlib.resources._adapters',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\resources\\_adapters.py',
+   'PYMODULE'),
+  ('importlib.resources._common',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\resources\\_common.py',
+   'PYMODULE'),
+  ('importlib.resources._itertools',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\resources\\_itertools.py',
+   'PYMODULE'),
+  ('importlib.resources._legacy',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\resources\\_legacy.py',
+   'PYMODULE'),
+  ('importlib.resources.abc',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\resources\\abc.py',
+   'PYMODULE'),
+  ('importlib.resources.readers',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\resources\\readers.py',
+   'PYMODULE'),
+  ('importlib.util',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\importlib\\util.py',
+   'PYMODULE'),
+  ('inspect',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\inspect.py',
+   'PYMODULE'),
+  ('ipaddress',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\ipaddress.py',
+   'PYMODULE'),
+  ('json',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\json\\__init__.py',
+   'PYMODULE'),
+  ('json.decoder',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\json\\decoder.py',
+   'PYMODULE'),
+  ('json.encoder',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\json\\encoder.py',
+   'PYMODULE'),
+  ('json.scanner',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\json\\scanner.py',
+   'PYMODULE'),
+  ('logging',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\logging\\__init__.py',
+   'PYMODULE'),
+  ('lzma', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\lzma.py', 'PYMODULE'),
+  ('numbers',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\numbers.py',
+   'PYMODULE'),
+  ('opcode', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\opcode.py', 'PYMODULE'),
+  ('pathlib',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\pathlib.py',
+   'PYMODULE'),
+  ('pickle', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\pickle.py', 'PYMODULE'),
+  ('pprint', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\pprint.py', 'PYMODULE'),
+  ('py_compile',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\py_compile.py',
+   'PYMODULE'),
+  ('quopri', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\quopri.py', 'PYMODULE'),
+  ('random', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\random.py', 'PYMODULE'),
+  ('selectors',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\selectors.py',
+   'PYMODULE'),
+  ('shutil', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\shutil.py', 'PYMODULE'),
+  ('signal', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\signal.py', 'PYMODULE'),
+  ('socket', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\socket.py', 'PYMODULE'),
+  ('statistics',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\statistics.py',
+   'PYMODULE'),
+  ('string', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\string.py', 'PYMODULE'),
+  ('stringprep',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\stringprep.py',
+   'PYMODULE'),
+  ('subprocess',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\subprocess.py',
+   'PYMODULE'),
+  ('tarfile',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\tarfile.py',
+   'PYMODULE'),
+  ('tempfile',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\tempfile.py',
+   'PYMODULE'),
+  ('textwrap',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\textwrap.py',
+   'PYMODULE'),
+  ('threading',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\threading.py',
+   'PYMODULE'),
+  ('token', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\token.py', 'PYMODULE'),
+  ('tokenize',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\tokenize.py',
+   'PYMODULE'),
+  ('tracemalloc',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\tracemalloc.py',
+   'PYMODULE'),
+  ('typing', 'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\typing.py', 'PYMODULE'),
+  ('urllib',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\urllib\\__init__.py',
+   'PYMODULE'),
+  ('urllib.parse',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\urllib\\parse.py',
+   'PYMODULE'),
+  ('zipfile',
+   'C:\\Users\\MK\\anaconda3\\envs\\WJ\\Lib\\zipfile.py',
+   'PYMODULE')])

installer/generated/build_pyinstaller/WhisperJAV/warn-WhisperJAV.txt

@@ -0,0 +1,26 @@
+
+This file lists modules PyInstaller was not able to find. This does not
+necessarily mean this module is required for running your program. Python and
+Python 3rd-party packages include a lot of conditional or optional modules. For
+example the module 'ntpath' only exists on Windows, whereas the module
+'posixpath' only exists on Posix systems.
+
+Types if import:
+* top-level: imported at the top-level - look at these first
+* conditional: imported within an if-statement
+* delayed: imported within a function
+* optional: imported within a try-except-statement
+
+IMPORTANT: Do NOT post this list to the issue-tracker. Use it as a basis for
+            tracking down the missing module yourself. Thanks!
+
+missing module named 'org.python' - imported by copy (optional)
+missing module named _frozen_importlib_external - imported by importlib._bootstrap (delayed), importlib (optional), importlib.abc (optional)
+excluded module named _frozen_importlib - imported by importlib (optional), importlib.abc (optional)
+missing module named org - imported by pickle (optional)
+missing module named posix - imported by os (conditional, optional), posixpath (optional), shutil (conditional), importlib._bootstrap_external (conditional)
+missing module named resource - imported by posix (top-level)
+missing module named pwd - imported by posixpath (delayed, conditional, optional), shutil (delayed, optional), tarfile (optional), pathlib (delayed, optional), subprocess (delayed, conditional, optional)
+missing module named grp - imported by shutil (delayed, optional), tarfile (optional), pathlib (delayed, optional), subprocess (delayed, conditional, optional)
+missing module named _posixsubprocess - imported by subprocess (conditional)
+missing module named fcntl - imported by subprocess (optional)

installer/generated/constraints_v1.7.5.txt

@@ -0,0 +1,109 @@
+# WhisperJAV v1.7.5 Dependency Constraints
+# ==========================================
+#
+# This file pins critical dependency versions to prevent conflicts caused by
+# speech enhancement packages (modelscope, clearvoice, bs-roformer, zipenhancer).
+#
+# These packages have aggressive dependency specifications that can break:
+# - numpy (modelscope wants <2.0, but we need >=2.0 for librosa 0.11+)
+# - datasets (modelscope breaks with datasets 4.x)
+# - huggingface-hub version compatibility
+# - librosa audio processing
+#
+# Usage: pip install -c constraints_v1.7.5.txt -r requirements_v1.7.5.txt
+#
+# IMPORTANT: This file is used during installation to prevent pip from
+# downgrading critical packages when installing speech enhancement backends.
+
+# =============================================================================
+# Core Scientific Stack (STRICT - Do NOT allow downgrades)
+# =============================================================================
+# numpy 2.x is required for librosa 0.11+ compatibility
+# modelscope/zipenhancer may try to install numpy<2.0 - BLOCK THIS
+numpy>=2.0
+
+# scipy 1.10.1+ required for audio processing compatibility
+scipy>=1.10.1
+
+# librosa 0.11+ supports numpy 2.x
+librosa>=0.11.0
+
+# =============================================================================
+# ML Ecosystem Stability (CRITICAL for modelscope compatibility)
+# =============================================================================
+# datasets 4.x breaks modelscope - MUST cap at <4.0
+datasets>=2.14.0,<4.0
+
+# pydantic 3.x has breaking API changes
+pydantic>=2.0,<3.0
+
+# HuggingFace ecosystem version alignment
+transformers>=4.40.0
+accelerate>=0.26.0
+huggingface-hub>=0.25.0
+
+# fsspec for filesystem operations (prevent unnecessary downgrade)
+fsspec>=2025.3.0
+
+# =============================================================================
+# System Utilities
+# =============================================================================
+# psutil for process management (clean subprocess termination)
+psutil>=5.9.0
+
+# pywebview for GUI framework
+pywebview>=5.0.0
+
+# =============================================================================
+# Audio/Video Processing
+# =============================================================================
+# Faster Whisper for efficient transcription
+faster-whisper>=1.1.0
+
+# silero-vad for voice activity detection
+silero-vad>=6.0
+
+# PySubtrans for subtitle translation
+PySubtrans>=0.7.0
+
+# OpenAI API for translation providers
+openai>=1.35.0
+
+# Google Gemini API for translation providers
+google-genai>=1.39.0
+
+# scikit-learn for semantic scene detection
+scikit-learn>=1.3.0
+
+# ONNX runtime for ZipEnhancer inference
+onnxruntime>=1.16.0
+
+# modelscope for ZipEnhancer speech enhancement
+modelscope>=1.20
+
+# =============================================================================
+# Package Conflict Notes
+# =============================================================================
+#
+# Known conflicts and mitigations:
+#
+# 1. modelscope + numpy:
+#    - modelscope may request numpy<2.0
+#    - We REQUIRE numpy>=2.0 for librosa 0.11+
+#    - Solution: Install numpy first, then modelscope with --no-deps flag
+#
+# 2. modelscope + datasets:
+#    - modelscope breaks with datasets>=4.0
+#    - Solution: Cap datasets<4.0 in constraints
+#
+# 3. clearvoice + numpy:
+#    - Original clearvoice requires numpy<2.0
+#    - Solution: Use forked version with numpy 2.x support
+#
+# 4. bs-roformer-infer:
+#    - Generally compatible but may pull old torchaudio
+#    - Solution: Install PyTorch/torchaudio first
+#
+# 5. zipenhancer (via modelscope):
+#    - Part of modelscope ecosystem
+#    - Solution: Install modelscope with constraints

installer/generated/construct_v1.7.5.yaml

@@ -0,0 +1,60 @@
+# Conda-Constructor Configuration for WhisperJAV v1.7.5
+# =========================================================
+#
+# This creates a Windows installer that:
+# - Installs minimal conda environment to %LOCALAPPDATA%\WhisperJAV
+# - Downloads PyTorch and dependencies during post-install
+# - Defers AI model downloads to first application run
+# - Checks for WebView2 and provides download link if missing
+#
+# Build command: constructor . --config construct_v1.7.5.yaml
+# Output: WhisperJAV-1.7.5-Windows-x86_64.exe
+
+name: WhisperJAV
+version: 1.7.5
+
+# Conda channels for package resolution
+channels:
+  - conda-forge
+  - defaults
+
+# Minimal base environment (heavy dependencies in post-install)
+specs:
+  - python=3.10.18  # Pinned for stability (tested version)
+  - pip             # Package installer
+  - git             # Required for GitHub installs
+  - ffmpeg>=6,<8    # Audio/video processing (v6 or v7 required, v8+ incompatible)
+  - menuinst        # Desktop/Start Menu shortcut creation
+
+# Install in dependency order for reliability
+install_in_dependency_order: true
+
+# Default installation path (user can override during install)
+default_prefix: '%LOCALAPPDATA%\WhisperJAV'
+
+# Post-installation script (runs after conda env setup)
+post_install: post_install_v1.7.5.bat
+
+# Files to include in installer package
+extra_files:
+  - post_install_v1.7.5.bat
+  - post_install_v1.7.5.py
+  - requirements_v1.7.5.txt
+  - constraints_v1.7.5.txt    # Version pinning for problematic packages
+  - WhisperJAV_Launcher_v1.7.5.py  # Script launcher (fallback if frozen exe missing)
+  # NOTE: WhisperJAV-GUI.exe is created by post_install from Scripts/whisperjav-gui.exe
+  # No need to include a frozen launcher - pip creates the entry point executable
+  - README_INSTALLER_v1.7.5.txt
+  - LICENSE.txt
+  - whisperjav_icon.ico
+  - whisperjav-1.7.5-py3-none-any.whl
+
+# Installer metadata
+license_file: LICENSE.txt
+welcome_image: whisperjav_icon.ico  # Optional: 164x314 PNG or ICO
+icon_image: whisperjav_icon.ico  # Installer executable icon
+nsis_template: custom_template_v1.7.5.nsi.tmpl  # Custom NSIS template with CreateShortcut
+
+# Installation options
+# keep_pkgs: false  # Delete package cache after install (saves space)
+# attempt_hardlinks: true  # Use hardlinks to save disk space

installer/generated/create_desktop_shortcut_v1.7.5.bat

@@ -0,0 +1,148 @@
+@echo off
+REM ========================================
+REM WhisperJAV v1.7.5 Desktop Shortcut Creator
+REM ========================================
+REM
+REM This script creates a Windows desktop shortcut that:
+REM - Launches pythonw.exe (no console window)
+REM - Runs whisperjav.webview_gui.main module
+REM - Uses the WhisperJAV icon
+REM - Sets working directory to installation root
+REM
+REM Uses PowerShell to create .lnk file via WScript.Shell COM object
+REM
+
+echo Creating WhisperJAV v1.7.5 desktop shortcut...
+
+REM Get the directory where this script is located (install root)
+set SCRIPT_DIR=%~dp0
+
+REM Shortcut configuration
+set SHORTCUT_NAME=WhisperJAV v1.7.5.lnk
+set TARGET_EXE=%SCRIPT_DIR%WhisperJAV-GUI.exe
+set PYTHONW=%SCRIPT_DIR%pythonw.exe
+set TARGET_ARGS=-m whisperjav.webview_gui.main
+set ICON=%SCRIPT_DIR%whisperjav_icon.ico
+
+REM Verify WhisperJAV-GUI.exe exists (preferred target)
+if exist "%TARGET_EXE%" (
+    echo Target: WhisperJAV-GUI.exe (standalone launcher)
+    set USE_LAUNCHER_EXE=1
+) else (
+    echo WARNING: WhisperJAV-GUI.exe not found, using pythonw fallback
+    set USE_LAUNCHER_EXE=0
+
+    REM Verify pythonw.exe exists as fallback
+    if not exist "%PYTHONW%" (
+        echo WARNING: pythonw.exe not found at: %PYTHONW%
+        echo Trying python.exe as fallback...
+        set PYTHONW=%SCRIPT_DIR%python.exe
+    )
+
+    if not exist "%PYTHONW%" (
+        echo ERROR: Neither WhisperJAV-GUI.exe nor python.exe found!
+        echo Cannot create shortcut.
+        exit /b 1
+    )
+)
+
+REM Verify icon exists
+if not exist "%ICON%" (
+    echo WARNING: Icon file not found at: %ICON%
+    echo Shortcut will be created without custom icon.
+    set ICON=
+)
+
+REM Create shortcut using PowerShell or VBScript fallback
+echo Icon: %ICON%
+echo Working Directory: %SCRIPT_DIR%
+echo.
+
+REM Try PowerShell first (modern Windows systems)
+where powershell >nul 2>nul
+if %ERRORLEVEL% EQU 0 (
+    echo Using PowerShell to create shortcut...
+    if "%USE_LAUNCHER_EXE%"=="1" (
+        REM Preferred: Point to WhisperJAV-GUI.exe
+        echo Shortcut target: %TARGET_EXE%
+        echo.
+        powershell -NoProfile -ExecutionPolicy Bypass -Command "$WshShell = New-Object -ComObject WScript.Shell; $Desktop = [Environment]::GetFolderPath('Desktop'); $ShortcutPath = Join-Path $Desktop '%SHORTCUT_NAME%'; $Shortcut = $WshShell.CreateShortcut($ShortcutPath); $Shortcut.TargetPath = '%TARGET_EXE%'; $Shortcut.WorkingDirectory = '%SCRIPT_DIR%'; if ('%ICON%' -ne '') { $Shortcut.IconLocation = '%ICON%' }; $Shortcut.Description = 'WhisperJAV v1.7.5 Japanese AV Subtitle Generator'; $Shortcut.Save()"
+    ) else (
+        REM Fallback: Point to pythonw.exe with module arguments
+        echo Shortcut target: %PYTHONW% %TARGET_ARGS%
+        echo.
+        powershell -NoProfile -ExecutionPolicy Bypass -Command "$WshShell = New-Object -ComObject WScript.Shell; $Desktop = [Environment]::GetFolderPath('Desktop'); $ShortcutPath = Join-Path $Desktop '%SHORTCUT_NAME%'; $Shortcut = $WshShell.CreateShortcut($ShortcutPath); $Shortcut.TargetPath = '%PYTHONW%'; $Shortcut.Arguments = '%TARGET_ARGS%'; $Shortcut.WorkingDirectory = '%SCRIPT_DIR%'; if ('%ICON%' -ne '') { $Shortcut.IconLocation = '%ICON%' }; $Shortcut.Description = 'WhisperJAV v1.7.5 Japanese AV Subtitle Generator'; $Shortcut.Save()"
+    )
+) else (
+    REM PowerShell not available - use VBScript fallback (works on all Windows)
+    echo PowerShell not found, using VBScript fallback...
+
+    REM Create temporary VBScript
+    set VBS_SCRIPT=%TEMP%\create_whisperjav_shortcut.vbs
+
+    if "%USE_LAUNCHER_EXE%"=="1" (
+        REM Preferred: Point to WhisperJAV-GUI.exe
+        echo Shortcut target: %TARGET_EXE%
+        echo.
+        (
+            echo Set WshShell = CreateObject^("WScript.Shell"^)
+            echo DesktopPath = WshShell.SpecialFolders^("Desktop"^)
+            echo Set Shortcut = WshShell.CreateShortcut^(DesktopPath ^& "\%SHORTCUT_NAME%"^)
+            echo Shortcut.TargetPath = "%TARGET_EXE%"
+            echo Shortcut.WorkingDirectory = "%SCRIPT_DIR%"
+            echo Shortcut.Description = "WhisperJAV v1.7.5 - Japanese AV Subtitle Generator"
+        ) > "%VBS_SCRIPT%"
+        REM Add icon line if icon exists
+        if not "%ICON%"=="" (
+            echo Shortcut.IconLocation = "%ICON%" >> "%VBS_SCRIPT%"
+        )
+        REM Complete VBScript
+        echo Shortcut.Save >> "%VBS_SCRIPT%"
+    ) else (
+        REM Fallback: Point to pythonw.exe with module arguments
+        echo Shortcut target: %PYTHONW% %TARGET_ARGS%
+        echo.
+        (
+            echo Set WshShell = CreateObject^("WScript.Shell"^)
+            echo DesktopPath = WshShell.SpecialFolders^("Desktop"^)
+            echo Set Shortcut = WshShell.CreateShortcut^(DesktopPath ^& "\%SHORTCUT_NAME%"^)
+            echo Shortcut.TargetPath = "%PYTHONW%"
+            echo Shortcut.Arguments = "%TARGET_ARGS%"
+            echo Shortcut.WorkingDirectory = "%SCRIPT_DIR%"
+            echo Shortcut.Description = "WhisperJAV v1.7.5 - Japanese AV Subtitle Generator"
+        ) > "%VBS_SCRIPT%"
+        REM Add icon line if icon exists
+        if not "%ICON%"=="" (
+            echo Shortcut.IconLocation = "%ICON%" >> "%VBS_SCRIPT%"
+        )
+        REM Complete VBScript
+        echo Shortcut.Save >> "%VBS_SCRIPT%"
+    )
+
+    REM Execute VBScript
+    cscript //NoLogo "%VBS_SCRIPT%"
+
+    REM Cleanup
+    del "%VBS_SCRIPT%" >nul 2>&1
+)
+
+REM Verify shortcut was created
+set DESKTOP_PATH=%USERPROFILE%\Desktop\%SHORTCUT_NAME%
+if exist "%DESKTOP_PATH%" (
+    echo.
+    echo ========================================
+    echo Desktop shortcut created successfully!
+    echo ========================================
+    echo Location: %DESKTOP_PATH%
+    echo.
+    exit /b 0
+) else (
+    echo.
+    echo ========================================
+    echo WARNING: Failed to create desktop shortcut
+    echo ========================================
+    echo You can manually create a shortcut to:
+    echo   %PYTHONW% %TARGET_ARGS%
+    echo.
+    exit /b 1
+)

installer/generated/post_install_v1.7.5.bat

@@ -0,0 +1,70 @@
+@echo off
+REM ========================================
+REM WhisperJAV v1.7.5 Post-Install Wrapper
+REM ========================================
+REM
+REM This batch file is called by conda-constructor after the base environment
+REM is created. Constructor places this script in pkgs/ and extra_files in the
+REM parent directory (installation root).
+REM
+REM This script launches the Python post-install script that handles:
+REM - Preflight checks (disk space, network, WebView2)
+REM - PyTorch installation (CUDA or CPU)
+REM - Python dependencies installation
+REM - WhisperJAV application installation
+REM - Desktop shortcut creation
+REM
+
+ECHO ========================================
+ECHO WhisperJAV v1.7.5 Post-Install
+ECHO ========================================
+ECHO.
+
+REM Verify Python executable exists in parent directory
+if not exist "%~dp0..\python.exe" (
+    ECHO ERROR: Cannot find python.exe
+    ECHO Expected location: %~dp0..\python.exe
+    ECHO Installation may be corrupted.
+    ECHO.
+    pause
+    exit /b 1
+)
+
+REM Verify Python post-install script exists in parent directory
+if not exist "%~dp0..\post_install_v1.7.5.py" (
+    ECHO ERROR: Cannot find post_install_v1.7.5.py
+    ECHO Expected location: %~dp0..\post_install_v1.7.5.py
+    ECHO Installation files may be missing.
+    ECHO.
+    pause
+    exit /b 1
+)
+
+ECHO Using Python: "%~dp0..\python.exe"
+ECHO Script location: "%~dp0..\post_install_v1.7.5.py"
+ECHO.
+
+REM Run the Python post-install script
+REM Both python.exe and the .py script are in the parent directory (../)
+"%~dp0..\python.exe" "%~dp0..\post_install_v1.7.5.py"
+
+REM Check exit code
+if %errorlevel% neq 0 (
+    ECHO.
+    ECHO ========================================
+    ECHO ERROR: Post-install script failed!
+    ECHO ========================================
+    ECHO Exit code: %errorlevel%
+    ECHO Check install_log_v1.7.5.txt for details
+    ECHO.
+    pause
+    exit /b %errorlevel%
+)
+
+ECHO.
+ECHO ========================================
+ECHO Post-Install completed successfully!
+ECHO ========================================
+ECHO.
+
+exit /b 0

installer/generated/requirements_v1.7.5.txt

@@ -0,0 +1,118 @@
+# WhisperJAV v1.7.5 Requirements
+# ================================
+#
+# This file lists all Python dependencies for WhisperJAV.
+# PyTorch is installed separately in post_install_v1.7.5.py
+# with CUDA version detection.
+#
+# IMPORTANT: Use with constraints file to prevent version conflicts:
+#   pip install -c constraints_v1.7.5.txt -r requirements_v1.7.5.txt
+#
+# Installation is performed in phases to ensure dependency compatibility.
+
+# ============================================================================
+# PHASE 1: Core Scientific Stack (Install FIRST)
+# ============================================================================
+# These packages MUST be installed first to establish version constraints
+# before any packages that might try to downgrade them.
+
+numpy>=2.0                # NumPy 2.x (required for librosa 0.11+)
+scipy>=1.10.1             # Signal processing
+librosa>=0.11.0           # Audio analysis (NumPy 2.x compatible)
+
+# ============================================================================
+# PHASE 2: Core Whisper Dependencies
+# ============================================================================
+# Git-based installs from main branches
+
+openai-whisper @ git+https://github.com/openai/whisper@main
+stable-ts @ git+https://github.com/meizhong986/stable-ts-fix-setup.git@main
+
+# Whisper variants
+faster-whisper>=1.1.0
+
+# ============================================================================
+# PHASE 3: Audio and Signal Processing
+# ============================================================================
+
+ffmpeg-python @ git+https://github.com/kkroening/ffmpeg-python.git  # PyPI tarball fails
+soundfile                 # Audio file I/O
+auditok                   # Voice Activity Detection (VAD)
+silero-vad>=6.0           # Silero VAD for scene detection
+ten-vad                   # TEN Framework VAD for speech segmentation
+pyloudnorm                # Audio normalization
+pydub                     # Audio manipulation
+
+# ============================================================================
+# PHASE 4: Speech Enhancement Backends (Required Functionality)
+# ============================================================================
+# CRITICAL: These packages have aggressive dependencies that can conflict.
+# Install with constraints.txt to prevent numpy/datasets downgrades.
+
+# ModelScope ecosystem (ZipEnhancer - lightweight SOTA)
+modelscope>=1.20
+addict                    # ModelScope dependency (dict with attribute access)
+datasets>=2.14.0,<4.0     # CRITICAL: 4.x breaks modelscope
+simplejson                # ModelScope dependency (JSON parsing)
+sortedcontainers          # ModelScope dependency (sorted collections)
+packaging                 # ModelScope dependency (version parsing)
+
+# ClearVoice denoising (48kHz) - NumPy 2.x compatible fork
+clearvoice @ git+https://github.com/meizhong986/ClearerVoice-Studio.git#subdirectory=clearvoice
+
+# BS-RoFormer vocal isolation (44.1kHz)
+bs-roformer-infer
+
+# ONNX runtime for ZipEnhancer ONNX mode
+onnxruntime>=1.16.0
+
+# ============================================================================
+# PHASE 5: GUI Framework (PyWebView)
+# ============================================================================
+
+pywebview>=5.0.0          # Modern web-based GUI framework
+pythonnet>=3.0            # Required for WebView2 backend on Windows
+pywin32>=305; platform_system=="Windows"  # Windows COM support for shortcuts
+
+# ============================================================================
+# PHASE 6: HuggingFace Transformers Pipeline
+# ============================================================================
+
+huggingface-hub>=0.25.0   # Hugging Face cache management (faster-whisper models)
+transformers>=4.40.0      # HuggingFace Transformers for ASR pipeline
+accelerate>=0.26.0        # Efficient model loading for Transformers
+hf_xet                    # Faster HuggingFace downloads (Xet Storage)
+
+# ============================================================================
+# PHASE 7: Subtitle Processing and Translation
+# ============================================================================
+
+pysrt                     # SRT file parsing
+srt                       # SRT utilities
+PySubtrans>=0.7.0         # AI-powered subtitle translation
+openai>=1.35.0            # Required for GPT provider in PySubtrans
+google-genai>=1.39.0      # Required for Gemini provider in PySubtrans
+
+# ============================================================================
+# PHASE 8: Utilities and Configuration
+# ============================================================================
+
+tqdm                      # Progress bars
+aiofiles                  # Async file I/O
+jsonschema                # Config validation
+Pillow                    # Image processing (GUI)
+colorama                  # Colored console output
+requests                  # HTTP requests
+regex                     # Advanced regular expressions
+psutil>=5.9.0             # Process management (clean subprocess cleanup)
+
+# Configuration system
+pydantic>=2.0,<3.0        # Data validation (used in config schemas)
+PyYAML>=6.0               # YAML config file parsing
+
+# ============================================================================
+# PHASE 9: Performance Optimization
+# ============================================================================
+
+numba                     # JIT compilation for speedups
+scikit-learn>=1.3.0       # Clustering for semantic scene detection

installer/generated/uninstall_v1.7.5.bat

@@ -0,0 +1,210 @@
+@echo off
+REM ===============================================================================
+REM WhisperJAV v1.7.5 Uninstaller
+REM ===============================================================================
+REM
+REM This script completely removes WhisperJAV v1.7.5 from your system:
+REM - Installation directory (all files)
+REM - Desktop shortcut
+REM - Optionally: User configuration
+REM - Optionally: Cached AI models (~3GB)
+REM
+REM This script should be run from the installation directory.
+REM
+
+SETLOCAL EnableDelayedExpansion
+
+echo ===============================================================================
+echo                      WhisperJAV v1.7.5 Uninstaller
+echo ===============================================================================
+echo.
+echo This will PERMANENTLY DELETE WhisperJAV v1.7.5 and all its files.
+echo.
+
+REM Get the installation directory (where this script is located)
+set INSTALL_DIR=%~dp0
+set INSTALL_DIR=%INSTALL_DIR:~0,-1%
+
+echo Installation Directory:
+echo   %INSTALL_DIR%
+echo.
+
+REM ===== Confirmation Prompt =====
+echo WARNING: This action cannot be undone!
+echo.
+set /p CONFIRM="Are you sure you want to uninstall WhisperJAV v1.7.5? (yes/no): "
+
+if /i not "%CONFIRM%"=="yes" (
+    echo.
+    echo Uninstallation canceled.
+    echo.
+    pause
+    exit /b 0
+)
+
+echo.
+echo ===============================================================================
+echo Starting uninstallation...
+echo ===============================================================================
+echo.
+
+REM Create uninstall log
+set UNINSTALL_LOG=%TEMP%\whisperjav_uninstall_v1.7.5.txt
+echo WhisperJAV v1.7.5 Uninstallation Log > "%UNINSTALL_LOG%"
+echo Uninstall started: %DATE% %TIME% >> "%UNINSTALL_LOG%"
+echo Installation directory: %INSTALL_DIR% >> "%UNINSTALL_LOG%"
+echo. >> "%UNINSTALL_LOG%"
+
+REM ===== Step 1: Remove Desktop Shortcut =====
+echo [Step 1/5] Removing desktop shortcut...
+echo [Step 1/5] Desktop shortcut >> "%UNINSTALL_LOG%"
+
+set SHORTCUT_PATH=%USERPROFILE%\Desktop\WhisperJAV v1.7.5.lnk
+
+if exist "%SHORTCUT_PATH%" (
+    del "%SHORTCUT_PATH%"
+    if exist "%SHORTCUT_PATH%" (
+        echo   WARNING: Could not delete shortcut
+        echo   WARNING: Shortcut deletion failed >> "%UNINSTALL_LOG%"
+    ) else (
+        echo   - Desktop shortcut removed
+        echo   - Desktop shortcut removed >> "%UNINSTALL_LOG%"
+    )
+) else (
+    echo   - Desktop shortcut not found (already removed)
+    echo   - Desktop shortcut not found >> "%UNINSTALL_LOG%"
+)
+echo.
+
+REM ===== Step 2: Optional - Remove User Configuration =====
+echo [Step 2/5] User configuration...
+echo [Step 2/5] User configuration >> "%UNINSTALL_LOG%"
+
+set CONFIG_FILE=%INSTALL_DIR%\whisperjav_config.json
+
+if exist "%CONFIG_FILE%" (
+    echo.
+    echo   Found user configuration file: whisperjav_config.json
+    echo   This contains your settings and preferences.
+    echo.
+    set /p DEL_CONFIG="   Delete configuration file? (y/N): "
+
+    if /i "!DEL_CONFIG!"=="y" (
+        del "%CONFIG_FILE%"
+        echo   - Configuration file deleted
+        echo   - Configuration file deleted >> "%UNINSTALL_LOG%"
+    ) else (
+        echo   - Configuration file kept
+        echo   - Configuration file kept (user choice) >> "%UNINSTALL_LOG%"
+    )
+) else (
+    echo   - No configuration file found
+    echo   - No configuration file >> "%UNINSTALL_LOG%"
+)
+echo.
+
+REM ===== Step 3: Optional - Remove Cached Models =====
+echo [Step 3/5] Cached AI models...
+echo [Step 3/5] Cached AI models >> "%UNINSTALL_LOG%"
+
+set MODEL_CACHE=%USERPROFILE%\.cache\whisper
+
+if exist "%MODEL_CACHE%" (
+    REM Calculate cache size
+    for /f "tokens=3" %%a in ('dir "%MODEL_CACHE%" /s /-c 2^>nul ^| findstr /c:"File(s)"') do set MODEL_SIZE=%%a
+    set /a MODEL_SIZE_MB=!MODEL_SIZE! / 1048576
+
+    echo.
+    echo   Found cached AI models in: %MODEL_CACHE%
+    echo   Cache size: ~!MODEL_SIZE_MB! MB
+    echo.
+    echo   Deleting these will free up disk space, but models will need
+    echo   to be re-downloaded if you reinstall WhisperJAV (~3GB, 5-10 min).
+    echo.
+    set /p DEL_MODELS="   Delete cached models? (y/N): "
+
+    if /i "!DEL_MODELS!"=="y" (
+        echo   - Deleting model cache... (this may take a minute)
+        rmdir /s /q "%MODEL_CACHE%"
+        if exist "%MODEL_CACHE%" (
+            echo   WARNING: Could not fully delete model cache
+            echo   WARNING: Model cache deletion failed >> "%UNINSTALL_LOG%"
+        ) else (
+            echo   - Model cache deleted (~!MODEL_SIZE_MB! MB freed)
+            echo   - Model cache deleted (~!MODEL_SIZE_MB! MB) >> "%UNINSTALL_LOG%"
+        )
+    ) else (
+        echo   - Model cache kept
+        echo   - Model cache kept (user choice) >> "%UNINSTALL_LOG%"
+    )
+) else (
+    echo   - No model cache found
+    echo   - No model cache found >> "%UNINSTALL_LOG%"
+)
+echo.
+
+REM ===== Step 4: Remove Installation Directory =====
+echo [Step 4/5] Removing installation directory...
+echo [Step 4/5] Installation directory >> "%UNINSTALL_LOG%"
+
+REM We need to delete the installation directory, but this script is running from it!
+REM Strategy: Create a temporary cleanup script that runs after this script exits
+
+set CLEANUP_SCRIPT=%TEMP%\whisperjav_cleanup_v1.7.5.bat
+
+echo @echo off > "%CLEANUP_SCRIPT%"
+echo REM Temporary cleanup script for WhisperJAV v1.7.5 >> "%CLEANUP_SCRIPT%"
+echo echo Removing installation directory... >> "%CLEANUP_SCRIPT%"
+echo timeout /t 2 /nobreak ^>nul >> "%CLEANUP_SCRIPT%"
+echo rmdir /s /q "%INSTALL_DIR%" >> "%CLEANUP_SCRIPT%"
+echo if exist "%INSTALL_DIR%" ( >> "%CLEANUP_SCRIPT%"
+echo     echo ERROR: Could not delete installation directory >> "%CLEANUP_SCRIPT%"
+echo     echo %INSTALL_DIR% >> "%CLEANUP_SCRIPT%"
+echo     echo. >> "%CLEANUP_SCRIPT%"
+echo     echo You may need to delete it manually. >> "%CLEANUP_SCRIPT%"
+echo     pause >> "%CLEANUP_SCRIPT%"
+echo ) else ( >> "%CLEANUP_SCRIPT%"
+echo     echo Installation directory removed successfully! >> "%CLEANUP_SCRIPT%"
+echo     echo. >> "%CLEANUP_SCRIPT%"
+echo     echo WhisperJAV v1.7.5 has been uninstalled. >> "%CLEANUP_SCRIPT%"
+echo     timeout /t 3 /nobreak ^>nul >> "%CLEANUP_SCRIPT%"
+echo ) >> "%CLEANUP_SCRIPT%"
+echo del "%%~f0" >> "%CLEANUP_SCRIPT%"
+
+echo   - Cleanup script created: %CLEANUP_SCRIPT%
+echo   - Cleanup script created >> "%UNINSTALL_LOG%"
+echo.
+
+REM ===== Step 5: Final Summary =====
+echo [Step 5/5] Uninstallation summary
+echo [Step 5/5] Uninstallation summary >> "%UNINSTALL_LOG%"
+
+echo.
+echo ===============================================================================
+echo                      Uninstallation Complete!
+echo ===============================================================================
+echo.
+echo The following items have been removed:
+echo   - Desktop shortcut: WhisperJAV v1.7.5.lnk
+if /i "%DEL_CONFIG%"=="y" echo   - User configuration file
+if /i "%DEL_MODELS%"=="y" echo   - Cached AI models (~!MODEL_SIZE_MB! MB)
+echo.
+echo The installation directory will be deleted when you close this window:
+echo   %INSTALL_DIR%
+echo.
+echo Uninstall log saved to:
+echo   %UNINSTALL_LOG%
+echo.
+echo ===============================================================================
+
+REM Log completion
+echo. >> "%UNINSTALL_LOG%"
+echo Uninstallation completed: %DATE% %TIME% >> "%UNINSTALL_LOG%"
+
+REM Launch cleanup script and exit
+echo.
+echo Press any key to finish uninstallation and close this window...
+pause >nul
+
+start "" "%CLEANUP_SCRIPT%"
+exit

installer/generated/validate_installer_v1.7.5.py

@@ -0,0 +1,506 @@
+#!/usr/bin/env python
+"""
+WhisperJAV v1.7.5 Installer Validation Script
+==============================================
+
+This script performs comprehensive pre-build validation checks to ensure
+all required files are present and correctly configured before running
+the conda-constructor build.
+
+Validation checks:
+1. All required files exist
+2. Version consistency across files
+3. Module paths are correct (no old GUI references)
+4. Requirements.txt encoding is valid (UTF-8)
+5. Icon file exists and is valid
+6. README and LICENSE present
+7. YAML syntax is valid
+
+Run this before building the installer to catch issues early.
+"""
+
+import os
+import sys
+import re
+from pathlib import Path
+
+# Configure UTF-8 encoding for Windows console
+if sys.platform == 'win32':
+    try:
+        import io
+        sys.stdout = io.TextIOWrapper(sys.stdout.buffer, encoding='utf-8', errors='replace')
+        sys.stderr = io.TextIOWrapper(sys.stderr.buffer, encoding='utf-8', errors='replace')
+    except Exception:
+        pass  # Fall back to default encoding if reconfiguration fails
+
+# ANSI color codes for better readability
+GREEN = '\033[92m'
+RED = '\033[91m'
+YELLOW = '\033[93m'
+BLUE = '\033[94m'
+RESET = '\033[0m'
+
+# Expected version
+EXPECTED_VERSION = "1.7.5"
+
+# Required files for v1.7.5 installer
+REQUIRED_FILES = [
+    "construct_v1.7.5.yaml",
+    "post_install_v1.7.5.bat",
+    "post_install_v1.7.5.py",
+    "requirements_v1.7.5.txt",
+    "constraints_v1.7.5.txt",    # Version pinning for problematic packages
+    "WhisperJAV_Launcher_v1.7.5.py",
+    "README_INSTALLER_v1.7.5.txt",
+    "LICENSE.txt",
+    "whisperjav_icon.ico",
+]
+
+# Optional files (warnings only, not errors)
+# NOTE: WhisperJAV-GUI.exe is created by post_install from Scripts/whisperjav-gui.exe
+OPTIONAL_FILES = []
+
+# Wheel file pattern - matches version with PEP 440 normalization
+# e.g., "1.7.0-beta" becomes "1.7.0b0" in wheel filename
+WHEEL_PATTERN = "whisperjav-1.7.5*.whl"
+WHEEL_PATTERN_FALLBACK = "whisperjav-*.whl"
+
+# Expected module path (no old Tkinter GUI references)
+CORRECT_MODULE_PATH = "whisperjav.webview_gui.main"
+OLD_MODULE_PATHS = [
+    "whisperjav.gui.whisperjav_gui",
+    "whisperjav.gui",
+]
+
+
+def print_header(title):
+    """Print a formatted section header"""
+    print()
+    print("=" * 80)
+    print(f"  {title}")
+    print("=" * 80)
+
+
+def print_check(description, passed, details=""):
+    """Print a check result with formatting"""
+    status = f"{GREEN}✓ PASS{RESET}" if passed else f"{RED}✗ FAIL{RESET}"
+    print(f"  [{status}] {description}")
+    if details:
+        print(f"          {details}")
+
+
+def check_file_exists(filepath):
+    """Check if a file exists"""
+    return Path(filepath).exists()
+
+
+def validate_required_files():
+    """Check that all required files exist"""
+    print_header("Phase 1: Required Files Check")
+
+    all_passed = True
+    for filename in REQUIRED_FILES:
+        exists = check_file_exists(filename)
+        print_check(filename, exists)
+        if not exists:
+            all_passed = False
+
+    # Check for wheel file using glob pattern (try version-specific first)
+    import glob
+    wheels = glob.glob(WHEEL_PATTERN)
+    if not wheels:
+        # Fall back to any wheel file
+        wheels = glob.glob(WHEEL_PATTERN_FALLBACK)
+
+    if wheels:
+        # Prefer the one matching our version
+        wheel_file = wheels[0]
+        for w in wheels:
+            if EXPECTED_VERSION.replace("-", "") in w or EXPECTED_VERSION in w:
+                wheel_file = w
+                break
+        print_check(f"Wheel file", True, f"Found: {wheel_file}")
+    else:
+        print_check("Wheel file", False, "No wheel file found matching pattern")
+        all_passed = False
+
+    return all_passed
+
+
+def validate_version_consistency():
+    """Check version consistency across files"""
+    print_header("Phase 2: Version Consistency Check")
+
+    all_passed = True
+    # Version regex supports both formats:
+    # - PEP 440: 1.7.0b0, 1.7.0a1, 1.7.0rc1 (no hyphen)
+    # - Display: 1.7.0-beta, 1.7.0-alpha (with hyphen)
+    version_pattern = r"(\d+\.\d+\.\d+(?:(?:a|b|rc)\d+)?(?:-[a-zA-Z0-9]+)?)"
+    version_files = [
+        ("construct_v1.7.5.yaml", rf"version:\s*['\"]?{version_pattern}"),
+        ("post_install_v1.7.5.py", rf"WhisperJAV v{version_pattern}"),
+        ("WhisperJAV_Launcher_v1.7.5.py", rf"v{version_pattern}"),
+        ("README_INSTALLER_v1.7.5.txt", rf"WhisperJAV v{version_pattern}"),
+    ]
+
+    for filename, pattern in version_files:
+        if not check_file_exists(filename):
+            print_check(f"Version in {filename}", False, "File not found")
+            all_passed = False
+            continue
+
+        try:
+            with open(filename, 'r', encoding='utf-8') as f:
+                content = f.read()
+
+            match = re.search(pattern, content)
+            if match:
+                version = match.group(1)
+                passed = (version == EXPECTED_VERSION)
+                details = f"Found: {version}" if not passed else ""
+                print_check(f"Version in {filename}: {version}", passed, details)
+                if not passed:
+                    all_passed = False
+            else:
+                print_check(f"Version in {filename}", False, "Version pattern not found")
+                all_passed = False
+
+        except Exception as e:
+            print_check(f"Version in {filename}", False, f"Error: {e}")
+            all_passed = False
+
+    return all_passed
+
+
+def validate_module_paths():
+    """Check that module paths reference the new PyWebView GUI, not old Tkinter GUI"""
+    print_header("Phase 3: Module Path Validation")
+
+    all_passed = True
+    # Files that contain GUI module references:
+    # - WhisperJAV_Launcher_v1.7.5.py: Script launcher
+    # - custom_template_v1.7.5.nsi.tmpl: NSIS shortcut creation (pythonw fallback)
+    # NOTE: post_install_v1.7.5.py removed - it only handles pip installation,
+    #       shortcuts are now handled by NSIS
+    files_to_check = [
+        "WhisperJAV_Launcher_v1.7.5.py",
+        "custom_template_v1.7.5.nsi.tmpl",
+    ]
+
+    for filename in files_to_check:
+        if not check_file_exists(filename):
+            print_check(f"Module paths in {filename}", False, "File not found")
+            all_passed = False
+            continue
+
+        try:
+            with open(filename, 'r', encoding='utf-8') as f:
+                content = f.read()
+
+            # Check for correct module path
+            has_correct_path = CORRECT_MODULE_PATH in content
+
+            # Check for old module paths (should not exist)
+            has_old_paths = any(old_path in content for old_path in OLD_MODULE_PATHS)
+
+            passed = has_correct_path and not has_old_paths
+
+            if passed:
+                print_check(f"Module paths in {filename}", True)
+            else:
+                if not has_correct_path:
+                    print_check(
+                        f"Module paths in {filename}",
+                        False,
+                        f"Missing correct path: {CORRECT_MODULE_PATH}"
+                    )
+                if has_old_paths:
+                    print_check(
+                        f"Module paths in {filename}",
+                        False,
+                        "Contains old Tkinter GUI references!"
+                    )
+                all_passed = False
+
+        except Exception as e:
+            print_check(f"Module paths in {filename}", False, f"Error: {e}")
+            all_passed = False
+
+    return all_passed
+
+
+def validate_requirements_encoding():
+    """Check that requirements.txt is valid UTF-8 without encoding issues"""
+    print_header("Phase 4: Requirements File Encoding Check")
+
+    filename = "requirements_v1.7.5.txt"
+
+    if not check_file_exists(filename):
+        print_check(f"{filename} encoding", False, "File not found")
+        return False
+
+    try:
+        with open(filename, 'r', encoding='utf-8') as f:
+            content = f.read()
+
+        # Check for weird spacing (sign of encoding issues)
+        if '  ' in content[:100]:  # Check first 100 chars for double spaces
+            print_check(
+                f"{filename} encoding",
+                False,
+                "Suspicious spacing detected - possible encoding issue"
+            )
+            return False
+
+        # Try to parse as requirements
+        lines = [line.strip() for line in content.split('\n') if line.strip() and not line.startswith('#')]
+
+        if len(lines) < 5:
+            print_check(
+                f"{filename} encoding",
+                False,
+                f"Too few packages ({len(lines)}) - file may be corrupted"
+            )
+            return False
+
+        print_check(f"{filename} encoding", True, f"{len(lines)} packages found")
+        return True
+
+    except UnicodeDecodeError as e:
+        print_check(f"{filename} encoding", False, f"UTF-8 decode error: {e}")
+        return False
+    except Exception as e:
+        print_check(f"{filename} encoding", False, f"Error: {e}")
+        return False
+
+
+def validate_optional_files():
+    """Check optional files (warnings only, not errors)"""
+    print_header("Phase 5: Optional Files Check")
+
+    # Optional files don't cause validation failure
+    for filename, description in OPTIONAL_FILES:
+        exists = check_file_exists(filename)
+        if exists:
+            size = Path(filename).stat().st_size
+            size_kb = size // 1024
+            print_check(f"{filename}", True, f"{size_kb} KB - {description}")
+        else:
+            # Print as warning (yellow), not error
+            print(f"  [{YELLOW}! WARN{RESET}] {filename}")
+            print(f"          Not found: {description}")
+            print(f"          (Will be created during build)")
+
+    # Always return True - optional files don't fail validation
+    return True
+
+
+def validate_assets():
+    """Check that icon and other assets exist and are valid"""
+    print_header("Phase 6: Asset File Validation")
+
+    all_passed = True
+
+    # Check icon file
+    icon_file = "whisperjav_icon.ico"
+    if check_file_exists(icon_file):
+        size = Path(icon_file).stat().st_size
+        if size > 1000:  # Should be at least 1KB for a valid icon
+            print_check(f"{icon_file}", True, f"{size} bytes")
+        else:
+            print_check(f"{icon_file}", False, f"File too small ({size} bytes)")
+            all_passed = False
+    else:
+        print_check(f"{icon_file}", False, "File not found")
+        all_passed = False
+
+    # Check LICENSE
+    license_file = "LICENSE.txt"
+    if check_file_exists(license_file):
+        size = Path(license_file).stat().st_size
+        if size > 100:  # Should be at least 100 bytes
+            print_check(f"{license_file}", True, f"{size} bytes")
+        else:
+            print_check(f"{license_file}", False, f"File too small ({size} bytes)")
+            all_passed = False
+    else:
+        print_check(f"{license_file}", False, "File not found")
+        all_passed = False
+
+    return all_passed
+
+
+def validate_constraints_file():
+    """Check that constraints file contains critical version pins"""
+    print_header("Phase 7: Constraints File Validation")
+
+    filename = "constraints_v1.7.5.txt"
+
+    if not check_file_exists(filename):
+        print_check(f"{filename}", False, "File not found")
+        return False
+
+    try:
+        with open(filename, 'r', encoding='utf-8') as f:
+            content = f.read()
+
+        # Check for critical version constraints
+        critical_constraints = [
+            ("numpy>=2.0", "NumPy 2.x for librosa compatibility"),
+            ("scipy>=1.10.1", "SciPy for audio processing"),
+            ("librosa>=0.11.0", "Librosa NumPy 2.x support"),
+            ("datasets>=2.14.0,<4.0", "Datasets cap for modelscope"),
+            ("pydantic>=2.0,<3.0", "Pydantic version cap"),
+        ]
+
+        all_found = True
+        for constraint, description in critical_constraints:
+            if constraint in content:
+                print_check(f"{constraint}", True, description)
+            else:
+                print_check(f"{constraint}", False, f"Missing: {description}")
+                all_found = False
+
+        return all_found
+
+    except Exception as e:
+        print_check(f"{filename}", False, f"Error: {e}")
+        return False
+
+
+def validate_yaml_syntax():
+    """Basic YAML syntax validation for construct file"""
+    print_header("Phase 8: YAML Syntax Validation")
+
+    filename = "construct_v1.7.5.yaml"
+
+    if not check_file_exists(filename):
+        print_check(f"{filename} syntax", False, "File not found")
+        return False
+
+    try:
+        import yaml
+        with open(filename, 'r', encoding='utf-8') as f:
+            data = yaml.safe_load(f)
+
+        # Check required keys
+        required_keys = ['name', 'version', 'channels', 'specs', 'post_install', 'extra_files']
+        missing_keys = [key for key in required_keys if key not in data]
+
+        if missing_keys:
+            print_check(
+                f"{filename} syntax",
+                False,
+                f"Missing required keys: {', '.join(missing_keys)}"
+            )
+            return False
+
+        # Verify version matches
+        if data['version'] != EXPECTED_VERSION:
+            print_check(
+                f"{filename} syntax",
+                False,
+                f"Version mismatch: {data['version']} != {EXPECTED_VERSION}"
+            )
+            return False
+
+        # Check extra_files exist (excluding optional files that are built during build process)
+        optional_filenames = [f for f, _ in OPTIONAL_FILES]
+        missing_files = [
+            f for f in data['extra_files']
+            if not check_file_exists(f) and f not in optional_filenames
+        ]
+        if missing_files:
+            print_check(
+                f"{filename} syntax",
+                False,
+                f"Extra files missing: {', '.join(missing_files)}"
+            )
+            return False
+
+        # Warn about optional files listed in extra_files that don't exist yet
+        optional_missing = [
+            f for f in data['extra_files']
+            if not check_file_exists(f) and f in optional_filenames
+        ]
+        if optional_missing:
+            print(f"  [{YELLOW}! NOTE{RESET}] Optional files in extra_files (built during build):")
+            for f in optional_missing:
+                print(f"          - {f}")
+
+        print_check(f"{filename} syntax", True, "YAML structure valid")
+        return True
+
+    except ImportError:
+        print_check(f"{filename} syntax", False, "PyYAML not installed (skipping)")
+        return True  # Non-fatal if yaml module not available
+    except Exception as e:
+        print_check(f"{filename} syntax", False, f"Error: {e}")
+        return False
+
+
+def main():
+    """Run all validation checks"""
+    print()
+    print("=" * 80)
+    print(" " * 20 + "WhisperJAV v1.7.5 Installer Validation")
+    print("=" * 80)
+    print()
+    print("Running pre-build validation checks...")
+
+    # Run all validation phases
+    results = []
+    results.append(("Required Files", validate_required_files()))
+    results.append(("Version Consistency", validate_version_consistency()))
+    results.append(("Module Paths", validate_module_paths()))
+    results.append(("Requirements Encoding", validate_requirements_encoding()))
+    results.append(("Optional Files", validate_optional_files()))
+    results.append(("Asset Files", validate_assets()))
+    results.append(("Constraints File", validate_constraints_file()))
+    results.append(("YAML Syntax", validate_yaml_syntax()))
+
+    # Print summary
+    print()
+    print("=" * 80)
+    print("  VALIDATION SUMMARY")
+    print("=" * 80)
+
+    all_passed = True
+    for phase, passed in results:
+        status = f"{GREEN}✓ PASS{RESET}" if passed else f"{RED}✗ FAIL{RESET}"
+        print(f"  {phase:.<50} [{status}]")
+        if not passed:
+            all_passed = False
+
+    print("=" * 80)
+
+    if all_passed:
+        print()
+        print(f"{GREEN}✓ All validation checks passed!{RESET}")
+        print()
+        print("The installer is ready to build.")
+        print("Run: build_installer_v1.7.5.bat")
+        print()
+        return 0
+    else:
+        print()
+        print(f"{RED}✗ Validation failed!{RESET}")
+        print()
+        print("Please fix the issues above before building the installer.")
+        print()
+        return 1
+
+
+if __name__ == "__main__":
+    try:
+        sys.exit(main())
+    except KeyboardInterrupt:
+        print()
+        print("Validation interrupted by user.")
+        sys.exit(1)
+    except Exception as e:
+        print()
+        print(f"FATAL ERROR: {e}")
+        import traceback
+        traceback.print_exc()
+        sys.exit(1)

installer/install_linux.sh

@@ -1,6 +1,6 @@
 #!/bin/bash
 # ==============================================================================
-# WhisperJAV Linux Installation Script
+# WhisperJAV Linux Installation Script (v1.7.5 Compatible)
 # ==============================================================================
 #
 # This script handles the staged installation of WhisperJAV on Linux systems,
@@ -11,20 +11,39 @@
 #   ./install_linux.sh
 #
 # Options:
-#   --cpu-only    Install CPU-only PyTorch (no CUDA)
-#   --no-speech-enhancement    Skip speech enhancement packages (clearvoice, etc.)
-#   --minimal     Install minimal version (no speech enhancement, no TEN VAD)
+#   --cpu-only              Install CPU-only PyTorch (no CUDA)
+#   --cuda121               Install PyTorch for CUDA 12.1
+#   --cuda124               Install PyTorch for CUDA 12.4
+#   --cuda126               Install PyTorch for CUDA 12.6
+#   --cuda128               Install PyTorch for CUDA 12.8 (default for driver 570+)
+#   --no-speech-enhancement Skip speech enhancement packages
+#   --minimal               Install minimal version (no speech enhancement, no TEN VAD)
+#   --dev                   Install in development/editable mode
 #
 # ==============================================================================
 
 set -e  # Exit on error
 
+# Initialize log file
+INSTALL_LOG="$(dirname "$0")/install_log_linux.txt"
+echo "" > "$INSTALL_LOG"
+
+# Logging function
+log() {
+    echo "[$(date '+%Y-%m-%d %H:%M:%S')] $1" >> "$INSTALL_LOG"
+}
+
 # Colors for output
 RED='\033[0;31m'
 GREEN='\033[0;32m'
 YELLOW='\033[1;33m'
 NC='\033[0m' # No Color
 
+log "============================================================"
+log "  WhisperJAV Linux Installation Script"
+log "  Started: $(date)"
+log "============================================================"
+
 echo "============================================================"
 echo "  WhisperJAV Linux Installation Script"
 echo "============================================================"
@@ -56,19 +75,49 @@ if ! command -v ffmpeg &> /dev/null; then
     echo "  Debian/Ubuntu: sudo apt-get install ffmpeg"
     echo "  Fedora/RHEL:   sudo dnf install ffmpeg"
     echo "  Arch Linux:    sudo pacman -S ffmpeg"
+    log "ERROR: FFmpeg not found"
     exit 1
 fi
 echo -e "${GREEN}FFmpeg found: $(ffmpeg -version | head -n1)${NC}"
+log "FFmpeg found"
+
+# Check for Git
+if ! command -v git &> /dev/null; then
+    echo -e "${RED}Error: Git is not installed.${NC}"
+    echo "Please install Git first:"
+    echo "  Debian/Ubuntu: sudo apt-get install git"
+    echo "  Fedora/RHEL:   sudo dnf install git"
+    echo "  Arch Linux:    sudo pacman -S git"
+    log "ERROR: Git not found"
+    exit 1
+fi
+echo -e "${GREEN}Git found${NC}"
+log "Git found"
 
 # Parse arguments
 CPU_ONLY=false
+CUDA_VERSION="auto"
 NO_SPEECH_ENHANCEMENT=false
 MINIMAL=false
+DEV_MODE=false
 
 for arg in "$@"; do
     case $arg in
         --cpu-only)
             CPU_ONLY=true
+            CUDA_VERSION="cpu"
+            ;;
+        --cuda121)
+            CUDA_VERSION="cuda121"
+            ;;
+        --cuda124)
+            CUDA_VERSION="cuda124"
+            ;;
+        --cuda126)
+            CUDA_VERSION="cuda126"
+            ;;
+        --cuda128)
+            CUDA_VERSION="cuda128"
             ;;
         --no-speech-enhancement)
             NO_SPEECH_ENHANCEMENT=true
@@ -77,129 +126,412 @@ for arg in "$@"; do
             MINIMAL=true
             NO_SPEECH_ENHANCEMENT=true
             ;;
+        --dev)
+            DEV_MODE=true
+            ;;
+        --help|-h)
+            echo ""
+            echo "WhisperJAV Linux Installation Script"
+            echo ""
+            echo "Usage: ./install_linux.sh [options]"
+            echo ""
+            echo "Options:"
+            echo "  --cpu-only              Install CPU-only PyTorch (no CUDA)"
+            echo "  --cuda121               Install PyTorch for CUDA 12.1"
+            echo "  --cuda124               Install PyTorch for CUDA 12.4"
+            echo "  --cuda126               Install PyTorch for CUDA 12.6"
+            echo "  --cuda128               Install PyTorch for CUDA 12.8 (default for driver 570+)"
+            echo "  --no-speech-enhancement Skip speech enhancement packages"
+            echo "  --minimal               Minimal install (no speech enhancement)"
+            echo "  --dev                   Install in development/editable mode"
+            echo "  --help, -h              Show this help message"
+            echo ""
+            echo "The script will auto-detect your GPU and select the appropriate"
+            echo "CUDA version. Use --cuda*** flags to override."
+            echo ""
+            exit 0
+            ;;
     esac
 done
 
 # Check Python version
 echo -e "${YELLOW}Checking Python version...${NC}"
+log "Checking Python version..."
 PYTHON_VERSION=$(python3 --version 2>&1 | awk '{print $2}')
 PYTHON_MAJOR=$(echo $PYTHON_VERSION | cut -d. -f1)
 PYTHON_MINOR=$(echo $PYTHON_VERSION | cut -d. -f2)
 
 if [ "$PYTHON_MAJOR" -lt 3 ] || ([ "$PYTHON_MAJOR" -eq 3 ] && [ "$PYTHON_MINOR" -lt 9 ]); then
     echo -e "${RED}Error: Python 3.9 or higher is required. Found: $PYTHON_VERSION${NC}"
+    log "ERROR: Python 3.9+ required. Found: $PYTHON_VERSION"
     exit 1
 fi
 
 if [ "$PYTHON_MINOR" -gt 12 ]; then
     echo -e "${YELLOW}Warning: Python 3.13+ may have compatibility issues with openai-whisper${NC}"
+    log "WARNING: Python 3.13+ detected"
 fi
 
 echo -e "${GREEN}Python $PYTHON_VERSION detected${NC}"
+log "Python $PYTHON_VERSION detected"
 echo ""
 
-# Auto-detect NVIDIA GPU
+# Auto-detect NVIDIA GPU and driver version
 if [ "$CPU_ONLY" = false ]; then
     echo -e "${YELLOW}Checking for NVIDIA GPU...${NC}"
+    log "Checking for NVIDIA GPU..."
     if command -v nvidia-smi &> /dev/null; then
         GPU_NAME=$(nvidia-smi --query-gpu=name --format=csv,noheader 2>/dev/null | head -n1)
+        DRIVER_VERSION=$(nvidia-smi --query-gpu=driver_version --format=csv,noheader 2>/dev/null | head -n1)
         if [ -n "$GPU_NAME" ]; then
             echo -e "${GREEN}NVIDIA GPU detected: $GPU_NAME${NC}"
+            echo -e "${GREEN}Driver version: $DRIVER_VERSION${NC}"
+            log "NVIDIA GPU detected: $GPU_NAME"
+            log "Driver version: $DRIVER_VERSION"
+
+            # Auto-select CUDA version based on driver if not specified
+            if [ "$CUDA_VERSION" = "auto" ]; then
+                DRIVER_MAJOR=$(echo "$DRIVER_VERSION" | cut -d. -f1)
+                if [ "$DRIVER_MAJOR" -ge 570 ] 2>/dev/null; then
+                    CUDA_VERSION="cuda128"
+                    echo -e "${GREEN}Auto-selecting CUDA 12.8 based on driver $DRIVER_VERSION${NC}"
+                    log "Auto-selecting CUDA 12.8"
+                elif [ "$DRIVER_MAJOR" -ge 560 ] 2>/dev/null; then
+                    CUDA_VERSION="cuda126"
+                    echo -e "${GREEN}Auto-selecting CUDA 12.6 based on driver $DRIVER_VERSION${NC}"
+                    log "Auto-selecting CUDA 12.6"
+                elif [ "$DRIVER_MAJOR" -ge 551 ] 2>/dev/null; then
+                    CUDA_VERSION="cuda124"
+                    echo -e "${GREEN}Auto-selecting CUDA 12.4 based on driver $DRIVER_VERSION${NC}"
+                    log "Auto-selecting CUDA 12.4"
+                else
+                    CUDA_VERSION="cuda121"
+                    echo -e "${GREEN}Using CUDA 12.1 for driver $DRIVER_VERSION${NC}"
+                    log "Using CUDA 12.1"
+                fi
+            fi
         else
             echo -e "${RED}nvidia-smi found but no GPU detected${NC}"
+            log "WARNING: nvidia-smi found but no GPU detected"
             CPU_ONLY=true
+            CUDA_VERSION="cpu"
         fi
     elif [ -f "/proc/driver/nvidia/version" ]; then
         echo -e "${GREEN}NVIDIA driver detected${NC}"
+        log "NVIDIA driver detected"
+        if [ "$CUDA_VERSION" = "auto" ]; then
+            CUDA_VERSION="cuda121"
+        fi
     else
         echo -e "${RED}No NVIDIA GPU detected!${NC}"
         echo -e "${YELLOW}Switching to CPU-only installation automatically.${NC}"
+        log "No NVIDIA GPU - switching to CPU-only"
         CPU_ONLY=true
+        CUDA_VERSION="cpu"
     fi
     echo ""
 fi
 
-# Upgrade pip
-echo -e "${YELLOW}Step 1/6: Upgrading pip...${NC}"
-python3 -m pip install --upgrade pip
-echo ""
-
-# Install PyTorch first (before anything else)
-echo -e "${YELLOW}Step 2/6: Installing PyTorch...${NC}"
+# Display configuration
+echo "============================================================"
+echo "  Installation Configuration"
+echo "============================================================"
 if [ "$CPU_ONLY" = true ]; then
-    echo "Installing CPU-only PyTorch..."
-    pip3 install torch torchaudio --index-url https://download.pytorch.org/whl/cpu
+    echo "  PyTorch: CPU-only"
+    log "Configuration: PyTorch=CPU-only"
 else
-    echo "Installing PyTorch with CUDA support..."
-    pip3 install torch torchaudio --index-url https://download.pytorch.org/whl/cu121
+    echo "  PyTorch: $CUDA_VERSION"
+    log "Configuration: PyTorch=$CUDA_VERSION"
 fi
+if [ "$NO_SPEECH_ENHANCEMENT" = true ]; then
+    echo "  Speech Enhancement: Disabled"
+    log "Configuration: Speech Enhancement=Disabled"
+else
+    echo "  Speech Enhancement: Enabled"
+    log "Configuration: Speech Enhancement=Enabled"
+fi
+if [ "$DEV_MODE" = true ]; then
+    echo "  Mode: Development (editable)"
+    log "Configuration: Mode=Development"
+else
+    echo "  Mode: Standard"
+    log "Configuration: Mode=Standard"
+fi
+echo "============================================================"
 echo ""
 
-# Install core dependencies that don't have conflicts
-echo -e "${YELLOW}Step 3/6: Installing core dependencies...${NC}"
-pip3 install numpy>=2.0 scipy>=1.10.1 librosa>=0.11.0
-pip3 install soundfile pydub tqdm colorama requests regex
+# ==============================================================================
+# Phase 1: Upgrade pip
+# ==============================================================================
+log ""
+log "============================================================"
+log "  Phase 1: Upgrade pip"
+log "============================================================"
+
+echo -e "${YELLOW}Step 1/7: Upgrading pip...${NC}"
+python3 -m pip install --upgrade pip
+log "pip upgraded"
+echo ""
+
+# ==============================================================================
+# Phase 2: Install PyTorch
+# ==============================================================================
+log ""
+log "============================================================"
+log "  Phase 2: PyTorch Installation"
+log "============================================================"
+
+echo -e "${YELLOW}Step 2/7: Installing PyTorch...${NC}"
+
+# Select PyTorch URL based on CUDA version
+case $CUDA_VERSION in
+    cpu)
+        TORCH_URL="https://download.pytorch.org/whl/cpu"
+        ;;
+    cuda121)
+        TORCH_URL="https://download.pytorch.org/whl/cu121"
+        ;;
+    cuda124)
+        TORCH_URL="https://download.pytorch.org/whl/cu124"
+        ;;
+    cuda126)
+        TORCH_URL="https://download.pytorch.org/whl/cu126"
+        ;;
+    cuda128)
+        TORCH_URL="https://download.pytorch.org/whl/cu128"
+        ;;
+    *)
+        TORCH_URL="https://download.pytorch.org/whl/cu121"
+        ;;
+esac
+
+echo "Installing PyTorch ($CUDA_VERSION)..."
+log "Installing PyTorch ($CUDA_VERSION) from $TORCH_URL..."
+pip3 install --progress-bar on torch torchaudio --index-url "$TORCH_URL"
+log "PyTorch installed"
+echo ""
+
+# ==============================================================================
+# Phase 3: Core Dependencies
+# ==============================================================================
+log ""
+log "============================================================"
+log "  Phase 3: Core Dependencies"
+log "============================================================"
+
+echo -e "${YELLOW}Step 3/7: Installing core dependencies...${NC}"
+
+# Phase 3.1: Core scientific stack (MUST install first)
+log "Phase 3.1: Installing scientific stack..."
+pip3 install "numpy>=2.0" "scipy>=1.10.1" "librosa>=0.11.0"
+
+# Phase 3.2: Audio and utility packages (including fsspec constraint)
+log "Phase 3.2: Installing audio/utility packages..."
+pip3 install --progress-bar on soundfile pydub tqdm colorama requests regex "psutil>=5.9.0" "fsspec>=2025.3.0"
+
+# Phase 3.3: Subtitle and async packages
+log "Phase 3.3: Installing subtitle/async packages..."
 pip3 install pysrt srt aiofiles jsonschema pyloudnorm
-pip3 install pydantic>=2.0 PyYAML>=6.0
+
+# Phase 3.4: Config and optimization packages
+log "Phase 3.4: Installing config packages..."
+pip3 install "pydantic>=2.0,<3.0" "PyYAML>=6.0" numba
+
+# Phase 3.5: Image packages (non-fatal)
+pip3 install Pillow || echo -e "${YELLOW}Warning: Pillow installation failed (non-fatal)${NC}"
+
+log "Core dependencies installed"
 echo ""
 
-# Install Whisper packages from git (avoid PyPI conflicts)
-echo -e "${YELLOW}Step 4/6: Installing Whisper packages...${NC}"
+# ==============================================================================
+# Phase 4: Whisper Packages
+# ==============================================================================
+log ""
+log "============================================================"
+log "  Phase 4: Whisper Packages"
+log "============================================================"
+
+echo -e "${YELLOW}Step 4/7: Installing Whisper packages...${NC}"
+log "Installing Whisper packages..."
+
 pip3 install git+https://github.com/openai/whisper@main
+log "openai-whisper installed"
+
 pip3 install git+https://github.com/meizhong986/stable-ts-fix-setup.git@main
-pip3 install faster-whisper>=1.1.0
+log "stable-ts installed"
+
+# ffmpeg-python: PyPI tarball fails, use git URL
+pip3 install git+https://github.com/kkroening/ffmpeg-python.git || {
+    echo -e "${YELLOW}Warning: ffmpeg-python from git failed, trying PyPI...${NC}"
+    log "WARNING: ffmpeg-python git failed, trying PyPI"
+    pip3 install ffmpeg-python
+}
+log "ffmpeg-python installed"
+
+pip3 install "faster-whisper>=1.1.0"
+log "faster-whisper installed"
 echo ""
 
-# Install optional packages
-echo -e "${YELLOW}Step 5/6: Installing optional packages...${NC}"
+# ==============================================================================
+# Phase 5: Optional Packages
+# ==============================================================================
+log ""
+log "============================================================"
+log "  Phase 5: Optional Packages"
+log "============================================================"
+
+echo -e "${YELLOW}Step 5/7: Installing optional packages...${NC}"
 
 # HuggingFace / Transformers
-pip3 install huggingface-hub>=0.25.0 transformers>=4.40.0 accelerate>=0.26.0
+log "Installing HuggingFace packages..."
+pip3 install "huggingface-hub>=0.25.0" "transformers>=4.40.0" "accelerate>=0.26.0"
+
+# hf_xet for faster HuggingFace downloads (optional)
+pip3 install hf_xet 2>/dev/null || log "Note: hf_xet not installed (optional)"
 
 # Translation
-pip3 install PySubtrans>=0.7.0 openai>=1.35.0 google-genai>=1.39.0
+log "Installing translation packages..."
+pip3 install "PySubtrans>=0.7.0" "openai>=1.35.0" "google-genai>=1.39.0"
 
 # VAD
-pip3 install silero-vad>=6.0 auditok
+log "Installing VAD packages..."
+pip3 install "silero-vad>=6.0" auditok
 
 if [ "$MINIMAL" = false ]; then
     pip3 install ten-vad || echo -e "${YELLOW}Warning: ten-vad installation failed (optional)${NC}"
+    pip3 install "scikit-learn>=1.3.0"
 fi
+log "Optional packages installed"
+echo ""
+
+# ==============================================================================
+# Phase 6: Speech Enhancement (optional but recommended)
+# ==============================================================================
+log ""
+log "============================================================"
+log "  Phase 6: Speech Enhancement"
+log "============================================================"
+
+if [ "$NO_SPEECH_ENHANCEMENT" = false ]; then
+    echo -e "${YELLOW}Step 6/7: Installing speech enhancement packages...${NC}"
+    echo "Note: These packages can be tricky. If installation fails, re-run with --no-speech-enhancement"
+    echo "      Speech enhancement improves transcription quality in noisy audio."
+    echo ""
+    log "Installing speech enhancement packages..."
+
+    # Install ModelScope dependencies first (CRITICAL: datasets must be <4.0)
+    log "Installing ModelScope dependencies..."
+    pip3 install addict simplejson sortedcontainers packaging
+    pip3 install "datasets>=2.14.0,<4.0" || {
+        echo -e "${YELLOW}Warning: datasets installation failed - modelscope may not work${NC}"
+        log "WARNING: datasets installation failed"
+    }
+
+    # Install ModelScope (ZipEnhancer SOTA speech enhancement)
+    echo "Installing ModelScope (ZipEnhancer)..."
+    log "Installing ModelScope..."
+    pip3 install "modelscope>=1.20" || {
+        echo -e "${YELLOW}Warning: modelscope installation failed (optional)${NC}"
+        log "WARNING: modelscope installation failed"
+    }
+
+    # Install ClearVoice from NumPy 2.x compatible fork
+    echo "Installing ClearVoice (48kHz denoising)..."
+    log "Installing ClearVoice..."
+    pip3 install "git+https://github.com/meizhong986/ClearerVoice-Studio.git#subdirectory=clearvoice" || {
+        echo -e "${YELLOW}Warning: clearvoice installation failed (optional)${NC}"
+        log "WARNING: clearvoice installation failed"
+    }
+
+    # BS-RoFormer (vocal isolation)
+    echo "Installing BS-RoFormer (vocal isolation)..."
+    log "Installing BS-RoFormer..."
+    pip3 install bs-roformer-infer || {
+        echo -e "${YELLOW}Warning: bs-roformer-infer installation failed (optional)${NC}"
+        log "WARNING: bs-roformer-infer installation failed"
+    }
+
+    # ONNX Runtime for ZipEnhancer ONNX mode
+    echo "Installing ONNX Runtime..."
+    pip3 install "onnxruntime>=1.16.0" || log "Note: onnxruntime installation failed (optional)"
+
+    log "Speech enhancement packages installed"
+else
+    echo -e "${YELLOW}Step 6/7: Skipping speech enhancement (--no-speech-enhancement)${NC}"
+    log "Skipping speech enhancement (--no-speech-enhancement)"
+fi
+echo ""
+
+# ==============================================================================
+# Phase 7: WhisperJAV Application
+# ==============================================================================
+log ""
+log "============================================================"
+log "  Phase 7: WhisperJAV Application"
+log "============================================================"
+
+echo -e "${YELLOW}Step 7/7: Installing WhisperJAV...${NC}"
+if [ "$DEV_MODE" = true ]; then
+    log "Installing WhisperJAV in development mode..."
+    pip3 install --no-deps -e .
+else
+    log "Installing WhisperJAV..."
+    pip3 install --no-deps git+https://github.com/meizhong986/whisperjav.git
+fi
+log "WhisperJAV installed"
+echo ""
+
+# ==============================================================================
+# Verification Phase
+# ==============================================================================
+log ""
+log "============================================================"
+log "  Verification Phase"
+log "============================================================"
+
+echo -e "${YELLOW}Verifying installation...${NC}"
+log "Verifying installation..."
+
+# Verify WhisperJAV
+python3 -c "import whisperjav; print(f'WhisperJAV {whisperjav.__version__} installed successfully!')" && {
+    echo -e "${GREEN}Installation complete!${NC}"
+    log "WhisperJAV verified successfully"
+} || {
+    echo -e "${RED}Installation may have issues. Check the output above.${NC}"
+    log "WARNING: WhisperJAV verification failed"
+}
+
+# Verify PyTorch CUDA
+python3 -c "import torch; cuda_status = 'ENABLED' if torch.cuda.is_available() else 'DISABLED'; print(f'CUDA acceleration: {cuda_status}')" 2>/dev/null
+
+# Verify critical packages
+echo ""
+echo "Verifying critical packages:"
+log "Verifying critical packages..."
+
+python3 -c "import numpy; print(f'  numpy: {numpy.__version__}')" 2>/dev/null || echo "  numpy: FAILED"
+python3 -c "import scipy; print(f'  scipy: {scipy.__version__}')" 2>/dev/null || echo "  scipy: FAILED"
+python3 -c "import librosa; print(f'  librosa: {librosa.__version__}')" 2>/dev/null || echo "  librosa: FAILED"
+python3 -c "import faster_whisper; print(f'  faster-whisper: {faster_whisper.__version__}')" 2>/dev/null || echo "  faster-whisper: FAILED"
+python3 -c "import transformers; print(f'  transformers: {transformers.__version__}')" 2>/dev/null || echo "  transformers: FAILED"
 
-# Speech Enhancement (optional, can cause conflicts)
 if [ "$NO_SPEECH_ENHANCEMENT" = false ]; then
     echo ""
-    echo -e "${YELLOW}Installing speech enhancement packages...${NC}"
-    echo "Note: These packages can be tricky. If installation fails, re-run with --no-speech-enhancement"
-
-    # Install modelscope and its dependencies first
-    pip3 install addict simplejson sortedcontainers packaging
-    pip3 install "datasets>=2.14.0,<4.0"
-    pip3 install modelscope>=1.20 || echo -e "${YELLOW}Warning: modelscope installation failed${NC}"
-
-    # Install clearvoice from fork (supports numpy 2.x)
-    pip3 install "git+https://github.com/meizhong986/ClearerVoice-Studio.git#subdirectory=clearvoice" || \
-        echo -e "${YELLOW}Warning: clearvoice installation failed (speech enhancement will be disabled)${NC}"
-
-    # BS-RoFormer
-    pip3 install bs-roformer-infer || echo -e "${YELLOW}Warning: bs-roformer-infer installation failed${NC}"
-
-    pip3 install onnxruntime>=1.16.0
+    echo "Speech enhancement packages:"
+    python3 -c "import modelscope; print(f'  modelscope: {modelscope.__version__}')" 2>/dev/null || echo "  modelscope: NOT INSTALLED"
+    python3 -c "import clearvoice; print('  clearvoice: installed')" 2>/dev/null || echo "  clearvoice: NOT INSTALLED"
 fi
-echo ""
-
-# Install WhisperJAV itself
-echo -e "${YELLOW}Step 6/6: Installing WhisperJAV...${NC}"
-pip3 install --no-deps git+https://github.com/meizhong986/whisperjav.git
-echo ""
-
-# Verify installation
-echo -e "${YELLOW}Verifying installation...${NC}"
-python3 -c "import whisperjav; print(f'WhisperJAV {whisperjav.__version__} installed successfully!')" && \
-    echo -e "${GREEN}Installation complete!${NC}" || \
-    echo -e "${RED}Installation may have issues. Check the output above.${NC}"
 
 echo ""
+
+# ==============================================================================
+# Summary
+# ==============================================================================
+log ""
+log "============================================================"
+log "  Installation Complete!"
+log "============================================================"
+
 echo "============================================================"
 echo "  Installation Summary"
 echo "============================================================"
@@ -213,3 +545,10 @@ echo ""
 echo "  For help:"
 echo "    whisperjav --help"
 echo ""
+echo "  Log file: $INSTALL_LOG"
+echo ""
+echo "  If you encounter issues with speech enhancement, re-run with:"
+echo "    ./install_linux.sh --no-speech-enhancement"
+echo ""
+
+log "Installation completed successfully at $(date)"

installer/install_windows.bat

@@ -17,9 +17,10 @@ REM
 REM Options:
 REM   --cpu-only              Install CPU-only PyTorch (no CUDA)
 REM   --cuda118               Install PyTorch for CUDA 11.8
-REM   --cuda121               Install PyTorch for CUDA 12.1 (default)
+REM   --cuda121               Install PyTorch for CUDA 12.1
 REM   --cuda124               Install PyTorch for CUDA 12.4
 REM   --cuda126               Install PyTorch for CUDA 12.6
+REM   --cuda128               Install PyTorch for CUDA 12.8 (default for driver 570+)
 REM   --no-speech-enhancement Skip speech enhancement packages
 REM   --minimal               Minimal install (transcription only)
 REM   --dev                   Install in development/editable mode
@@ -58,6 +59,7 @@ if /i "%~1"=="--cuda118" set "CUDA_VERSION=cuda118"
 if /i "%~1"=="--cuda121" set "CUDA_VERSION=cuda121"
 if /i "%~1"=="--cuda124" set "CUDA_VERSION=cuda124"
 if /i "%~1"=="--cuda126" set "CUDA_VERSION=cuda126"
+if /i "%~1"=="--cuda128" set "CUDA_VERSION=cuda128"
 if /i "%~1"=="--no-speech-enhancement" set "NO_SPEECH_ENHANCEMENT=1"
 if /i "%~1"=="--minimal" (
     set "MINIMAL=1"
@@ -220,12 +222,17 @@ if "%CPU_ONLY%"=="0" (
         REM Auto-select CUDA version based on driver if user didn't specify
         if "%CUDA_VERSION%"=="cuda121" (
             REM Check if driver supports newer CUDA
+            REM Driver 570+ supports CUDA 12.8
             REM Driver 560+ supports CUDA 12.6
             REM Driver 551+ supports CUDA 12.4
             REM Driver 531+ supports CUDA 12.1
             for /f "tokens=1 delims=." %%d in ("!DRIVER_VERSION!") do set "DRIVER_MAJOR=%%d"
             if defined DRIVER_MAJOR (
-                if !DRIVER_MAJOR! GEQ 560 (
+                if !DRIVER_MAJOR! GEQ 570 (
+                    echo Auto-selecting CUDA 12.8 based on driver !DRIVER_VERSION!
+                    call :log "Auto-selecting CUDA 12.8 based on driver"
+                    set "CUDA_VERSION=cuda128"
+                ) else if !DRIVER_MAJOR! GEQ 560 (
                     echo Auto-selecting CUDA 12.6 based on driver !DRIVER_VERSION!
                     call :log "Auto-selecting CUDA 12.6 based on driver"
                     set "CUDA_VERSION=cuda126"
@@ -286,6 +293,10 @@ if "%CUDA_VERSION%"=="cuda126" (
     set "TORCH_URL=https://download.pytorch.org/whl/cu126"
     set "TORCH_PACKAGES=torch torchaudio"
 )
+if "%CUDA_VERSION%"=="cuda128" (
+    set "TORCH_URL=https://download.pytorch.org/whl/cu128"
+    set "TORCH_PACKAGES=torch torchaudio"
+)
 
 REM Display configuration
 echo.
@@ -350,7 +361,7 @@ if errorlevel 1 (
 echo.
 
 REM ============================================================
-REM Phase 4: Core Dependencies
+REM Phase 4: Core Dependencies (with constraints)
 REM ============================================================
 call :log ""
 call :log "============================================================"
@@ -361,23 +372,31 @@ echo ============================================================
 echo   [Step 3/7] Installing core dependencies
 echo ============================================================
 
+REM Phase 4.1: Core scientific stack (MUST install first to establish versions)
+call :log "Phase 4.1: Installing core scientific stack..."
 call :run_pip_with_retry "install numpy>=2.0 scipy>=1.10.1 librosa>=0.11.0" "Install scientific packages"
 if errorlevel 1 goto :install_failed
 
-call :run_pip_with_retry "install soundfile pydub tqdm colorama requests regex" "Install audio/utility packages"
+REM Phase 4.2: Audio and utility packages
+call :log "Phase 4.2: Installing audio/utility packages..."
+call :run_pip_with_retry "install --progress-bar on soundfile pydub tqdm colorama requests regex psutil>=5.9.0 fsspec>=2025.3.0" "Install audio/utility packages"
 if errorlevel 1 goto :install_failed
 
+REM Phase 4.3: Subtitle and async packages
+call :log "Phase 4.3: Installing subtitle/async packages..."
 call :run_pip_with_retry "install pysrt srt aiofiles jsonschema pyloudnorm" "Install subtitle/async packages"
 if errorlevel 1 goto :install_failed
 
+REM Phase 4.4: Config and optimization packages
+call :log "Phase 4.4: Installing config packages..."
 call :run_pip_with_retry "install pydantic>=2.0,<3.0 PyYAML>=6.0 numba" "Install config packages"
 if errorlevel 1 goto :install_failed
 
-REM Additional packages from requirements_v1.7.4.txt
-call :run_pip_with_retry "install Pillow matplotlib" "Install image/plotting packages"
+REM Phase 4.5: Image/plotting packages (non-fatal)
+call :run_pip_with_retry "install Pillow" "Install image packages"
 if errorlevel 1 (
-    echo WARNING: Pillow/matplotlib installation failed (non-fatal)
-    call :log "WARNING: Pillow/matplotlib installation failed"
+    echo WARNING: Pillow installation failed (non-fatal)
+    call :log "WARNING: Pillow installation failed"
 )
 echo.
 
@@ -405,10 +424,12 @@ if errorlevel 1 (
     exit /b 1
 )
 
-call :run_pip_with_retry "install git+https://github.com/kkroening/ffmpeg-python.git" "Install ffmpeg-python"
+REM ffmpeg-python: PyPI tarball fails, use git URL
+call :run_pip_with_retry "install git+https://github.com/kkroening/ffmpeg-python.git" "Install ffmpeg-python (git)"
 if errorlevel 1 (
     echo WARNING: ffmpeg-python from git failed, trying PyPI version...
-    call :run_pip_with_retry "install ffmpeg-python" "Install ffmpeg-python (PyPI)"
+    call :log "WARNING: ffmpeg-python git failed, trying PyPI..."
+    call :run_pip_with_retry "install ffmpeg-python" "Install ffmpeg-python (PyPI fallback)"
 )
 
 call :run_pip_with_retry "install faster-whisper>=1.1.0" "Install faster-whisper"
@@ -438,6 +459,12 @@ if errorlevel 1 (
     call :log "WARNING: HuggingFace packages failed"
 )
 
+REM hf_xet for faster HuggingFace downloads (optional)
+python -m pip install hf_xet 2>nul
+if errorlevel 1 (
+    call :log "Note: hf_xet not installed (optional, faster HF downloads)"
+)
+
 REM Translation
 call :log "Installing translation packages..."
 call :run_pip_with_retry "install PySubtrans>=0.7.0 openai>=1.35.0 google-genai>=1.39.0" "Install translation packages"
@@ -471,40 +498,58 @@ if "%MINIMAL%"=="0" (
 )
 echo.
 
-REM Speech Enhancement (optional)
+REM Speech Enhancement (optional but recommended)
 if "%NO_SPEECH_ENHANCEMENT%"=="0" (
     echo ============================================================
     echo   [Step 6/7] Installing speech enhancement packages
     echo ============================================================
     echo Note: These packages can be tricky. Failures here are non-fatal.
+    echo       Speech enhancement improves transcription quality in noisy audio.
     echo.
     call :log "Installing speech enhancement packages..."
 
+    REM Install ModelScope dependencies first (CRITICAL: datasets must be <4.0)
+    call :log "Installing ModelScope dependencies..."
     python -m pip install addict simplejson sortedcontainers packaging 2>nul
     python -m pip install "datasets>=2.14.0,<4.0" 2>nul
+    if errorlevel 1 (
+        echo WARNING: datasets installation failed - modelscope may not work
+        call :log "WARNING: datasets installation failed"
+    )
 
-    echo Installing ModelScope...
+    REM Install ModelScope (ZipEnhancer SOTA speech enhancement)
+    echo Installing ModelScope ^(ZipEnhancer^)...
+    call :log "Installing ModelScope..."
     python -m pip install "modelscope>=1.20" 2>nul
     if errorlevel 1 (
         echo WARNING: modelscope installation failed ^(optional^)
         call :log "WARNING: modelscope installation failed"
     )
 
-    echo Installing ClearVoice...
+    REM Install ClearVoice from NumPy 2.x compatible fork
+    echo Installing ClearVoice ^(48kHz denoising^)...
+    call :log "Installing ClearVoice..."
     python -m pip install "git+https://github.com/meizhong986/ClearerVoice-Studio.git#subdirectory=clearvoice" 2>nul
     if errorlevel 1 (
         echo WARNING: clearvoice installation failed ^(optional^)
         call :log "WARNING: clearvoice installation failed"
     )
 
-    echo Installing BS-RoFormer...
+    REM Install BS-RoFormer (vocal isolation)
+    echo Installing BS-RoFormer ^(vocal isolation^)...
+    call :log "Installing BS-RoFormer..."
     python -m pip install bs-roformer-infer 2>nul
     if errorlevel 1 (
         echo WARNING: bs-roformer-infer installation failed ^(optional^)
         call :log "WARNING: bs-roformer-infer installation failed"
     )
 
+    REM ONNX Runtime for ZipEnhancer ONNX mode
+    echo Installing ONNX Runtime...
     python -m pip install "onnxruntime>=1.16.0" 2>nul
+    if errorlevel 1 (
+        call :log "Note: onnxruntime installation failed (optional)"
+    )
     echo.
 ) else (
     echo ============================================================
@@ -546,7 +591,7 @@ if errorlevel 1 (
 echo.
 
 REM ============================================================
-REM Verification and Summary
+REM Verification Phase: Check Critical Dependencies
 REM ============================================================
 call :log ""
 call :log "============================================================"
@@ -556,9 +601,11 @@ call :log "============================================================"
 echo ============================================================
 echo   Verifying Installation
 echo ============================================================
+
+REM Verify WhisperJAV
 python -c "import whisperjav; print(f'WhisperJAV {whisperjav.__version__} installed successfully!')"
 if errorlevel 1 (
-    echo WARNING: Could not verify installation
+    echo WARNING: Could not verify WhisperJAV installation
     call :log "WARNING: WhisperJAV verification failed"
 ) else (
     call :log "WhisperJAV verified successfully"
@@ -567,6 +614,24 @@ if errorlevel 1 (
 REM Verify PyTorch CUDA status
 python -c "import torch; cuda_status = 'ENABLED' if torch.cuda.is_available() else 'DISABLED'; print(f'CUDA acceleration: {cuda_status}')" 2>nul
 
+REM Verify critical packages
+echo.
+echo Verifying critical packages:
+call :log "Verifying critical packages..."
+
+python -c "import numpy; print(f'  numpy: {numpy.__version__}')" 2>nul || echo   numpy: FAILED
+python -c "import scipy; print(f'  scipy: {scipy.__version__}')" 2>nul || echo   scipy: FAILED
+python -c "import librosa; print(f'  librosa: {librosa.__version__}')" 2>nul || echo   librosa: FAILED
+python -c "import faster_whisper; print(f'  faster-whisper: {faster_whisper.__version__}')" 2>nul || echo   faster-whisper: FAILED
+python -c "import transformers; print(f'  transformers: {transformers.__version__}')" 2>nul || echo   transformers: FAILED
+
+if "%NO_SPEECH_ENHANCEMENT%"=="0" (
+    echo.
+    echo Speech enhancement packages:
+    python -c "import modelscope; print(f'  modelscope: {modelscope.__version__}')" 2>nul || echo   modelscope: NOT INSTALLED
+    python -c "import clearvoice; print('  clearvoice: installed')" 2>nul || echo   clearvoice: NOT INSTALLED
+)
+
 echo.
 
 REM Summary
@@ -674,9 +739,10 @@ echo.
 echo Options:
 echo   --cpu-only              Install CPU-only PyTorch ^(no CUDA^)
 echo   --cuda118               Install PyTorch for CUDA 11.8
-echo   --cuda121               Install PyTorch for CUDA 12.1 ^(default^)
+echo   --cuda121               Install PyTorch for CUDA 12.1
 echo   --cuda124               Install PyTorch for CUDA 12.4
 echo   --cuda126               Install PyTorch for CUDA 12.6
+echo   --cuda128               Install PyTorch for CUDA 12.8 ^(default for driver 570+^)
 echo   --no-speech-enhancement Skip speech enhancement packages
 echo   --minimal               Minimal install ^(transcription only^)
 echo   --dev                   Install in development/editable mode
@@ -685,9 +751,11 @@ echo.
 echo Examples:
 echo   install_windows.bat                    # Standard install with auto CUDA detection
 echo   install_windows.bat --cpu-only         # CPU-only install
+echo   install_windows.bat --cuda128          # Force CUDA 12.8
 echo   install_windows.bat --minimal --dev    # Minimal dev install
 echo.
 echo The script will auto-detect your GPU and select the appropriate
-echo CUDA version. Use --cuda*** flags to override.
+echo CUDA version ^(12.8 for driver 570+, 12.6 for 560+, etc.^).
+echo Use --cuda*** flags to override.
 echo.
 exit /b 0

tests/test_compute_type_auto.py

@@ -0,0 +1,565 @@
+#!/usr/bin/env python3
+"""
+Unit tests for compute_type="auto" strategy.
+
+Tests the CTranslate2 compute_type delegation implemented in:
+- whisperjav/config/resolver_v3.py
+- whisperjav/modules/faster_whisper_pro_asr.py
+- whisperjav/modules/kotoba_faster_whisper_asr.py
+- whisperjav/modules/speech_segmentation/backends/whisper_vad.py
+
+Background:
+- CTranslate2's "auto" selects optimal compute_type based on device capabilities
+- Automatically handles RTX 50XX Blackwell (sm120) which doesn't support INT8
+- See: https://github.com/OpenNMT/CTranslate2/issues/1865
+
+Run with: pytest tests/test_compute_type_auto.py -v
+"""
+
+import pytest
+from unittest.mock import patch, MagicMock
+
+
+# =============================================================================
+# Resolver Tests
+# =============================================================================
+
+class TestResolverComputeType:
+    """Tests for resolver_v3.py compute_type selection."""
+
+    def test_ctranslate2_providers_return_auto(self):
+        """Test that CTranslate2 providers get compute_type='auto'."""
+        from whisperjav.config.resolver_v3 import _get_compute_type_for_device
+
+        # CTranslate2 providers should return "auto" regardless of device
+        assert _get_compute_type_for_device("cuda", "faster_whisper") == "auto"
+        assert _get_compute_type_for_device("cpu", "faster_whisper") == "auto"
+        assert _get_compute_type_for_device("cuda", "kotoba_faster_whisper") == "auto"
+        assert _get_compute_type_for_device("cpu", "kotoba_faster_whisper") == "auto"
+
+    def test_pytorch_providers_return_explicit_types(self):
+        """Test that PyTorch providers get explicit compute_type values."""
+        from whisperjav.config.resolver_v3 import _get_compute_type_for_device
+
+        # PyTorch providers should return explicit types
+        assert _get_compute_type_for_device("cuda", "openai_whisper") == "float16"
+        assert _get_compute_type_for_device("mps", "openai_whisper") == "float16"
+        assert _get_compute_type_for_device("cpu", "openai_whisper") == "float32"
+        assert _get_compute_type_for_device("cuda", "stable_ts") == "float16"
+        assert _get_compute_type_for_device("cpu", "stable_ts") == "float32"
+
+    def test_ctranslate2_providers_set(self):
+        """Test that CTRANSLATE2_PROVIDERS set is correctly defined."""
+        from whisperjav.config.resolver_v3 import CTRANSLATE2_PROVIDERS
+
+        assert "faster_whisper" in CTRANSLATE2_PROVIDERS
+        assert "kotoba_faster_whisper" in CTRANSLATE2_PROVIDERS
+        # PyTorch providers should NOT be in this set
+        assert "openai_whisper" not in CTRANSLATE2_PROVIDERS
+        assert "stable_ts" not in CTRANSLATE2_PROVIDERS
+
+
+class TestResolverIntegration:
+    """Integration tests for resolver with compute_type='auto'."""
+
+    @pytest.fixture
+    def mock_device_detector(self):
+        """Mock device detector to return specific device."""
+        with patch('whisperjav.config.resolver_v3.get_best_device') as mock:
+            yield mock
+
+    def test_resolve_config_faster_whisper_cuda(self, mock_device_detector):
+        """Test that faster_whisper config has compute_type='auto' on CUDA."""
+        mock_device_detector.return_value = "cuda"
+
+        # Mock the component registries
+        with patch('whisperjav.config.resolver_v3.get_asr_registry') as mock_asr:
+            with patch('whisperjav.config.resolver_v3.get_vad_registry') as mock_vad:
+                # Setup mock ASR component
+                mock_component = MagicMock()
+                mock_component.provider = "faster_whisper"
+                mock_component.model_id = "large-v3"
+                mock_component.supported_tasks = ["transcribe"]
+                mock_preset = MagicMock()
+                mock_preset.model_dump.return_value = {"language": "ja", "task": "transcribe"}
+                mock_component.get_preset.return_value = mock_preset
+
+                mock_asr.return_value = {"faster_whisper": mock_component}
+                mock_vad.return_value = {}
+
+                from whisperjav.config.resolver_v3 import resolve_config_v3
+                config = resolve_config_v3(asr="faster_whisper", vad="none")
+
+                assert config["model"]["compute_type"] == "auto"
+
+    def test_resolve_config_faster_whisper_cpu(self, mock_device_detector):
+        """Test that faster_whisper config has compute_type='auto' on CPU."""
+        mock_device_detector.return_value = "cpu"
+
+        with patch('whisperjav.config.resolver_v3.get_asr_registry') as mock_asr:
+            with patch('whisperjav.config.resolver_v3.get_vad_registry') as mock_vad:
+                mock_component = MagicMock()
+                mock_component.provider = "faster_whisper"
+                mock_component.model_id = "large-v3"
+                mock_component.supported_tasks = ["transcribe"]
+                mock_preset = MagicMock()
+                mock_preset.model_dump.return_value = {"language": "ja", "task": "transcribe"}
+                mock_component.get_preset.return_value = mock_preset
+
+                mock_asr.return_value = {"faster_whisper": mock_component}
+                mock_vad.return_value = {}
+
+                from whisperjav.config.resolver_v3 import resolve_config_v3
+                config = resolve_config_v3(asr="faster_whisper", vad="none")
+
+                # Should be "auto" even on CPU - CTranslate2 handles selection
+                assert config["model"]["compute_type"] == "auto"
+
+
+# =============================================================================
+# ASR Module Tests
+# =============================================================================
+
+class TestFasterWhisperProASR:
+    """Tests for FasterWhisperProASR compute_type handling."""
+
+    def test_uses_compute_type_from_config(self):
+        """Test that compute_type is taken from model_config."""
+        # We can't easily test the full initialization without models
+        # but we can verify the default behavior
+        pass  # Tested via integration tests
+
+    def test_default_compute_type_is_auto(self):
+        """Test that default compute_type is 'auto'."""
+        # This is verified by checking the resolver returns "auto"
+        from whisperjav.config.resolver_v3 import _get_compute_type_for_device
+        assert _get_compute_type_for_device("cuda", "faster_whisper") == "auto"
+
+
+class TestKotobaFasterWhisperASR:
+    """Tests for KotobaFasterWhisperASR compute_type handling."""
+
+    def test_default_compute_type_is_auto(self):
+        """Test that default compute_type is 'auto'."""
+        from whisperjav.config.resolver_v3 import _get_compute_type_for_device
+        assert _get_compute_type_for_device("cuda", "kotoba_faster_whisper") == "auto"
+
+
+class TestWhisperVAD:
+    """Tests for WhisperVAD compute_type handling."""
+
+    def test_default_compute_type_is_auto(self):
+        """Test that default compute_type parameter is 'auto'."""
+        import inspect
+        from whisperjav.modules.speech_segmentation.backends.whisper_vad import WhisperVadSpeechSegmenter
+
+        sig = inspect.signature(WhisperVadSpeechSegmenter.__init__)
+        compute_type_param = sig.parameters.get('compute_type')
+
+        assert compute_type_param is not None
+        assert compute_type_param.default == "auto"
+
+
+# =============================================================================
+# CTranslate2 Auto Selection Logic Tests
+# =============================================================================
+
+class TestCTranslate2AutoBehavior:
+    """Tests documenting expected CTranslate2 'auto' behavior."""
+
+    def test_auto_documented_behavior(self):
+        """
+        Document expected CTranslate2 'auto' behavior.
+
+        According to CTranslate2 documentation:
+        - "auto" selects the fastest supported compute type
+        - GPU (CC ≥8.0): All types supported, likely selects int8_float16
+        - GPU (CC 6.1-7.0): int8/float16 optimized
+        - CPU: int8/int16/float32 based on instruction sets
+        - Blackwell sm120: Disables INT8 (PR #1937)
+
+        This test documents the expected behavior rather than testing it,
+        as the actual behavior depends on CTranslate2 internals.
+        """
+        expected_behaviors = {
+            "RTX 40XX (Ampere/Ada)": "int8_float16 (fastest with tensor cores)",
+            "RTX 50XX (Blackwell sm120)": "float16 (INT8 disabled by PR #1937)",
+            "CPU with AVX2": "int8 (quantized, fast)",
+            "CPU without AVX2": "float32 (fallback)",
+        }
+
+        # This is documentation, not a test
+        assert len(expected_behaviors) > 0
+
+
+class TestBackwardsCompatibility:
+    """Tests for backwards compatibility with explicit compute_type."""
+
+    def test_explicit_compute_type_still_works(self):
+        """Test that explicit compute_type values are still respected."""
+        # When users explicitly set compute_type, it should be used
+        # The "auto" is only the default from resolver
+
+        from whisperjav.config.resolver_v3 import _get_compute_type_for_device
+
+        # Resolver returns "auto" for CTranslate2 providers
+        result = _get_compute_type_for_device("cuda", "faster_whisper")
+        assert result == "auto"
+
+        # But ASR modules accept explicit overrides via model_config
+        # This is tested via the actual code path:
+        # model_config.get("compute_type", "auto")
+        # If user provides {"compute_type": "int8"}, that value is used
+
+
+# =============================================================================
+# CLI Device/Compute-Type Override Tests
+# =============================================================================
+
+class TestCLIDeviceOverride:
+    """Tests for --device and --compute-type CLI argument flow."""
+
+    @pytest.fixture
+    def mock_registries(self):
+        """Mock component registries for testing."""
+        with patch('whisperjav.config.resolver_v3.get_asr_registry') as mock_asr:
+            with patch('whisperjav.config.resolver_v3.get_vad_registry') as mock_vad:
+                # Setup mock ASR component
+                mock_component = MagicMock()
+                mock_component.provider = "faster_whisper"
+                mock_component.model_id = "large-v3"
+                mock_component.supported_tasks = ["transcribe"]
+                mock_preset = MagicMock()
+                mock_preset.model_dump.return_value = {"language": "ja", "task": "transcribe"}
+                mock_component.get_preset.return_value = mock_preset
+
+                mock_asr.return_value = {"faster_whisper": mock_component}
+                mock_vad.return_value = {}
+                yield mock_asr, mock_vad
+
+    def test_device_override_cuda(self, mock_registries):
+        """Test that explicit device='cuda' overrides auto-detection."""
+        with patch('whisperjav.config.resolver_v3.get_best_device') as mock_device:
+            mock_device.return_value = "cpu"  # Would auto-detect CPU
+
+            from whisperjav.config.resolver_v3 import resolve_config_v3
+            config = resolve_config_v3(
+                asr="faster_whisper",
+                vad="none",
+                device="cuda",  # Override to CUDA
+            )
+
+            assert config["model"]["device"] == "cuda"
+            # get_best_device should NOT be called when device is explicitly provided
+            # (Actually it's still called for safety, but result is ignored)
+
+    def test_device_override_cpu(self, mock_registries):
+        """Test that explicit device='cpu' overrides auto-detection."""
+        with patch('whisperjav.config.resolver_v3.get_best_device') as mock_device:
+            mock_device.return_value = "cuda"  # Would auto-detect CUDA
+
+            from whisperjav.config.resolver_v3 import resolve_config_v3
+            config = resolve_config_v3(
+                asr="faster_whisper",
+                vad="none",
+                device="cpu",  # Override to CPU
+            )
+
+            assert config["model"]["device"] == "cpu"
+
+    def test_device_auto_uses_detection(self, mock_registries):
+        """Test that device='auto' uses auto-detection."""
+        with patch('whisperjav.config.resolver_v3.get_best_device') as mock_device:
+            mock_device.return_value = "cuda"
+
+            from whisperjav.config.resolver_v3 import resolve_config_v3
+            config = resolve_config_v3(
+                asr="faster_whisper",
+                vad="none",
+                device="auto",  # Explicit auto
+            )
+
+            assert config["model"]["device"] == "cuda"
+            mock_device.assert_called_once()
+
+    def test_device_none_uses_detection(self, mock_registries):
+        """Test that device=None (default) uses auto-detection."""
+        with patch('whisperjav.config.resolver_v3.get_best_device') as mock_device:
+            mock_device.return_value = "cpu"
+
+            from whisperjav.config.resolver_v3 import resolve_config_v3
+            config = resolve_config_v3(
+                asr="faster_whisper",
+                vad="none",
+                # device not specified = None
+            )
+
+            assert config["model"]["device"] == "cpu"
+            mock_device.assert_called_once()
+
+    def test_compute_type_override_int8(self, mock_registries):
+        """Test that explicit compute_type='int8' overrides auto."""
+        with patch('whisperjav.config.resolver_v3.get_best_device') as mock_device:
+            mock_device.return_value = "cuda"
+
+            from whisperjav.config.resolver_v3 import resolve_config_v3
+            config = resolve_config_v3(
+                asr="faster_whisper",
+                vad="none",
+                compute_type="int8",  # Override
+            )
+
+            assert config["model"]["compute_type"] == "int8"
+
+    def test_compute_type_override_float16(self, mock_registries):
+        """Test that explicit compute_type='float16' overrides auto."""
+        with patch('whisperjav.config.resolver_v3.get_best_device') as mock_device:
+            mock_device.return_value = "cuda"
+
+            from whisperjav.config.resolver_v3 import resolve_config_v3
+            config = resolve_config_v3(
+                asr="faster_whisper",
+                vad="none",
+                compute_type="float16",
+            )
+
+            assert config["model"]["compute_type"] == "float16"
+
+    def test_compute_type_override_int8_float16(self, mock_registries):
+        """Test that explicit compute_type='int8_float16' works."""
+        with patch('whisperjav.config.resolver_v3.get_best_device') as mock_device:
+            mock_device.return_value = "cuda"
+
+            from whisperjav.config.resolver_v3 import resolve_config_v3
+            config = resolve_config_v3(
+                asr="faster_whisper",
+                vad="none",
+                compute_type="int8_float16",
+            )
+
+            assert config["model"]["compute_type"] == "int8_float16"
+
+    def test_compute_type_auto_uses_provider_default(self, mock_registries):
+        """Test that compute_type='auto' uses provider-specific default."""
+        with patch('whisperjav.config.resolver_v3.get_best_device') as mock_device:
+            mock_device.return_value = "cuda"
+
+            from whisperjav.config.resolver_v3 import resolve_config_v3
+            config = resolve_config_v3(
+                asr="faster_whisper",
+                vad="none",
+                compute_type="auto",  # Explicit auto
+            )
+
+            # CTranslate2 provider should get "auto"
+            assert config["model"]["compute_type"] == "auto"
+
+    def test_compute_type_none_uses_provider_default(self, mock_registries):
+        """Test that compute_type=None uses provider-specific default."""
+        with patch('whisperjav.config.resolver_v3.get_best_device') as mock_device:
+            mock_device.return_value = "cuda"
+
+            from whisperjav.config.resolver_v3 import resolve_config_v3
+            config = resolve_config_v3(
+                asr="faster_whisper",
+                vad="none",
+                # compute_type not specified = None
+            )
+
+            assert config["model"]["compute_type"] == "auto"
+
+    def test_combined_device_and_compute_type_override(self, mock_registries):
+        """Test that both device and compute_type can be overridden together."""
+        with patch('whisperjav.config.resolver_v3.get_best_device') as mock_device:
+            mock_device.return_value = "cuda"  # Would detect CUDA
+
+            from whisperjav.config.resolver_v3 import resolve_config_v3
+            config = resolve_config_v3(
+                asr="faster_whisper",
+                vad="none",
+                device="cpu",  # Force CPU
+                compute_type="float32",  # Force float32
+            )
+
+            assert config["model"]["device"] == "cpu"
+            assert config["model"]["compute_type"] == "float32"
+
+
+class TestLegacyPipelineDeviceOverride:
+    """Tests for device/compute_type override through legacy.py."""
+
+    @pytest.fixture
+    def mock_resolve_v3(self):
+        """Mock resolve_config_v3 to verify parameters passed."""
+        with patch('whisperjav.config.legacy.resolve_config_v3') as mock:
+            # Return a minimal valid config
+            mock.return_value = {
+                'asr_name': 'faster_whisper',
+                'vad_name': 'none',
+                'sensitivity_name': 'balanced',
+                'task': 'transcribe',
+                'language': 'ja',
+                'model': {
+                    'provider': 'faster_whisper',
+                    'model_name': 'large-v3',
+                    'device': 'cuda',
+                    'compute_type': 'auto',
+                    'supported_tasks': ['transcribe']
+                },
+                'params': {'asr': {}, 'vad': {}},
+                'features': {},
+            }
+            yield mock
+
+    def test_resolve_legacy_pipeline_passes_device(self, mock_resolve_v3):
+        """Test that resolve_legacy_pipeline passes device parameter."""
+        from whisperjav.config.legacy import resolve_legacy_pipeline
+
+        resolve_legacy_pipeline(
+            pipeline_name="balanced",
+            sensitivity="aggressive",
+            device="cpu",
+        )
+
+        # Verify resolve_config_v3 was called with device parameter
+        mock_resolve_v3.assert_called_once()
+        call_kwargs = mock_resolve_v3.call_args[1]
+        assert call_kwargs.get('device') == "cpu"
+
+    def test_resolve_legacy_pipeline_passes_compute_type(self, mock_resolve_v3):
+        """Test that resolve_legacy_pipeline passes compute_type parameter."""
+        from whisperjav.config.legacy import resolve_legacy_pipeline
+
+        resolve_legacy_pipeline(
+            pipeline_name="balanced",
+            sensitivity="balanced",
+            compute_type="int8_float16",
+        )
+
+        call_kwargs = mock_resolve_v3.call_args[1]
+        assert call_kwargs.get('compute_type') == "int8_float16"
+
+    def test_resolve_legacy_pipeline_passes_both(self, mock_resolve_v3):
+        """Test that resolve_legacy_pipeline passes both device and compute_type."""
+        from whisperjav.config.legacy import resolve_legacy_pipeline
+
+        resolve_legacy_pipeline(
+            pipeline_name="faster",
+            device="cuda",
+            compute_type="float16",
+        )
+
+        call_kwargs = mock_resolve_v3.call_args[1]
+        assert call_kwargs.get('device') == "cuda"
+        assert call_kwargs.get('compute_type') == "float16"
+
+    def test_resolve_legacy_pipeline_none_defaults(self, mock_resolve_v3):
+        """Test that resolve_legacy_pipeline passes None when not specified."""
+        from whisperjav.config.legacy import resolve_legacy_pipeline
+
+        resolve_legacy_pipeline(
+            pipeline_name="balanced",
+            # device and compute_type not specified
+        )
+
+        call_kwargs = mock_resolve_v3.call_args[1]
+        assert call_kwargs.get('device') is None
+        assert call_kwargs.get('compute_type') is None
+
+
+class TestEnsembleConfigDeviceOverride:
+    """Tests for device/compute_type override through ensemble config."""
+
+    @pytest.fixture
+    def mock_resolve_v3(self):
+        """Mock resolve_config_v3 for ensemble config tests."""
+        with patch('whisperjav.config.legacy.resolve_config_v3') as mock:
+            mock.return_value = {
+                'asr_name': 'faster_whisper',
+                'vad_name': 'none',
+                'sensitivity_name': 'balanced',
+                'task': 'transcribe',
+                'language': 'ja',
+                'model': {
+                    'provider': 'faster_whisper',
+                    'model_name': 'large-v3',
+                    'device': 'cuda',
+                    'compute_type': 'auto',
+                    'supported_tasks': ['transcribe']
+                },
+                'params': {'asr': {}, 'vad': {}},
+                'features': {},
+            }
+            yield mock
+
+    def test_resolve_ensemble_config_passes_device(self, mock_resolve_v3):
+        """Test that resolve_ensemble_config passes device parameter."""
+        from whisperjav.config.legacy import resolve_ensemble_config
+
+        resolve_ensemble_config(
+            asr="faster_whisper",
+            device="cpu",
+        )
+
+        call_kwargs = mock_resolve_v3.call_args[1]
+        assert call_kwargs.get('device') == "cpu"
+
+    def test_resolve_ensemble_config_passes_compute_type(self, mock_resolve_v3):
+        """Test that resolve_ensemble_config passes compute_type parameter."""
+        from whisperjav.config.legacy import resolve_ensemble_config
+
+        resolve_ensemble_config(
+            asr="faster_whisper",
+            compute_type="int8",
+        )
+
+        call_kwargs = mock_resolve_v3.call_args[1]
+        assert call_kwargs.get('compute_type') == "int8"
+
+
+class TestCLIArgumentParsing:
+    """Tests for CLI argument parsing in main.py."""
+
+    def test_device_argument_choices(self):
+        """Test that --device accepts valid choices."""
+        from whisperjav.main import parse_arguments
+        import sys
+
+        # Test with valid choices
+        for device in ["auto", "cuda", "cpu"]:
+            with patch.object(sys, 'argv', ['whisperjav', 'test.mp4', f'--device={device}']):
+                args = parse_arguments()
+                assert args.device == device
+
+    def test_compute_type_argument_choices(self):
+        """Test that --compute-type accepts valid choices."""
+        from whisperjav.main import parse_arguments
+        import sys
+
+        valid_types = ["auto", "float16", "float32", "int8", "int8_float16", "int8_float32"]
+        for compute_type in valid_types:
+            with patch.object(sys, 'argv', ['whisperjav', 'test.mp4', f'--compute-type={compute_type}']):
+                args = parse_arguments()
+                assert args.compute_type == compute_type
+
+    def test_device_default_is_none(self):
+        """Test that --device defaults to None (auto-detect)."""
+        from whisperjav.main import parse_arguments
+        import sys
+
+        with patch.object(sys, 'argv', ['whisperjav', 'test.mp4']):
+            args = parse_arguments()
+            assert args.device is None
+
+    def test_compute_type_default_is_none(self):
+        """Test that --compute-type defaults to None (auto)."""
+        from whisperjav.main import parse_arguments
+        import sys
+
+        with patch.object(sys, 'argv', ['whisperjav', 'test.mp4']):
+            args = parse_arguments()
+            assert args.compute_type is None
+
+
+if __name__ == "__main__":
+    pytest.main([__file__, "-v"])

tests/test_update_check_api.py

@@ -0,0 +1,661 @@
+#!/usr/bin/env python3
+"""
+Unit tests for WhisperJAV GUI Update Check API.
+
+Tests the update check modal feature including:
+- check_for_updates() API method
+- dismiss_update_notification() API method
+- get_dismissed_update() API method
+- open_url() API method
+- start_update() API method
+
+Run with: pytest tests/test_update_check_api.py -v
+"""
+
+import json
+import os
+import pytest
+import subprocess
+import sys
+import tempfile
+from datetime import datetime, timedelta
+from pathlib import Path
+from typing import NamedTuple
+from unittest.mock import patch, MagicMock, Mock
+
+
+# =============================================================================
+# Mock Classes to simulate version_checker types
+# =============================================================================
+
+class MockVersionInfo(NamedTuple):
+    """Mock VersionInfo to avoid importing actual module."""
+    version: str
+    release_url: str
+    release_notes: str
+    published_at: str
+    is_prerelease: bool
+    installer_url: str = None
+
+
+class MockUpdateCheckResult(NamedTuple):
+    """Mock UpdateCheckResult to avoid importing actual module."""
+    update_available: bool
+    current_version: str
+    latest_version: str
+    version_info: MockVersionInfo = None
+    from_cache: bool = False
+    error: str = None
+
+
+# =============================================================================
+# Fixtures
+# =============================================================================
+
+@pytest.fixture
+def temp_cache_dir(tmp_path):
+    """Create a temporary cache directory."""
+    cache_dir = tmp_path / ".whisperjav_cache"
+    cache_dir.mkdir()
+    return cache_dir
+
+
+@pytest.fixture
+def mock_api(tmp_path, monkeypatch):
+    """
+    Create a mock API instance with patched dependencies.
+
+    This patches the environment to use temp directories and
+    avoids importing heavy dependencies like webview.
+    """
+    # Patch LOCALAPPDATA to use temp directory
+    monkeypatch.setenv('LOCALAPPDATA', str(tmp_path))
+
+    # Create a minimal API class for testing
+    # We import just the methods we need to test
+    from whisperjav.webview_gui.api import WhisperJAVAPI
+
+    # Create API instance (may need window mock)
+    with patch('webview.create_window'):
+        api = WhisperJAVAPI.__new__(WhisperJAVAPI)
+        api.window = None
+        api._process = None
+        api._output_queue = None
+        api._log_thread = None
+        api._status = {"running": False}
+
+    return api
+
+
+@pytest.fixture
+def sample_update_result():
+    """Sample update check result with update available."""
+    return MockUpdateCheckResult(
+        update_available=True,
+        current_version="1.7.5",
+        latest_version="1.8.0",
+        version_info=MockVersionInfo(
+            version="1.8.0",
+            release_url="https://github.com/meizhong986/whisperjav/releases/tag/v1.8.0",
+            release_notes="## What's New\n- Feature A\n- Feature B\n- Bug fixes",
+            published_at="2025-01-15T10:00:00Z",
+            is_prerelease=False,
+            installer_url="https://github.com/meizhong986/whisperjav/releases/download/v1.8.0/WhisperJAV-1.8.0-Windows-x86_64.exe"
+        ),
+        from_cache=False,
+        error=None
+    )
+
+
+@pytest.fixture
+def sample_no_update_result():
+    """Sample update check result with no update available."""
+    return MockUpdateCheckResult(
+        update_available=False,
+        current_version="1.8.0",
+        latest_version="1.8.0",
+        version_info=None,
+        from_cache=True,
+        error=None
+    )
+
+
+@pytest.fixture
+def sample_error_result():
+    """Sample update check result with error."""
+    return MockUpdateCheckResult(
+        update_available=False,
+        current_version="1.7.5",
+        latest_version=None,
+        version_info=None,
+        from_cache=False,
+        error="Network error: Unable to connect to GitHub"
+    )
+
+
+# =============================================================================
+# check_for_updates() Tests
+# =============================================================================
+
+class TestCheckForUpdates:
+    """Tests for check_for_updates() API method."""
+
+    def test_update_available_returns_correct_structure(self, mock_api, sample_update_result):
+        """Test that update available response has correct structure."""
+        with patch('whisperjav.version_checker.check_for_updates', return_value=sample_update_result):
+            with patch('whisperjav.version_checker.get_update_notification_level', return_value='minor'):
+                result = mock_api.check_for_updates(force=True)
+
+        assert result["success"] is True
+        assert result["update_available"] is True
+        assert result["current_version"] == "1.7.5"
+        assert result["latest_version"] == "1.8.0"
+        assert result["notification_level"] == "minor"
+        assert result["release_url"] is not None
+        assert result["release_notes"] is not None
+        assert result["from_cache"] is False
+
+    def test_no_update_available(self, mock_api, sample_no_update_result):
+        """Test response when no update is available."""
+        with patch('whisperjav.version_checker.check_for_updates', return_value=sample_no_update_result):
+            result = mock_api.check_for_updates(force=False)
+
+        assert result["success"] is True
+        assert result["update_available"] is False
+        assert result["current_version"] == "1.8.0"
+        assert result["from_cache"] is True
+        # No notification_level, release_url, release_notes when no update
+        assert "notification_level" not in result or result.get("notification_level") is None
+
+    def test_error_handling(self, mock_api, sample_error_result):
+        """Test response when update check fails."""
+        with patch('whisperjav.version_checker.check_for_updates', return_value=sample_error_result):
+            result = mock_api.check_for_updates(force=True)
+
+        assert result["success"] is True  # Method succeeded, but check had error
+        assert result["update_available"] is False
+        assert "error" in result
+        assert "Network error" in result["error"]
+
+    def test_exception_handling(self, mock_api):
+        """Test that exceptions are caught and returned as error."""
+        with patch('whisperjav.version_checker.check_for_updates', side_effect=Exception("Import failed")):
+            result = mock_api.check_for_updates(force=True)
+
+        assert result["success"] is False
+        assert "Import failed" in result["error"]
+        assert result["update_available"] is False
+
+    def test_force_bypasses_cache(self, mock_api, sample_update_result):
+        """Test that force=True bypasses cache."""
+        mock_check = Mock(return_value=sample_update_result)
+        with patch('whisperjav.version_checker.check_for_updates', mock_check):
+            with patch('whisperjav.version_checker.get_update_notification_level', return_value='minor'):
+                mock_api.check_for_updates(force=True)
+
+        # Verify force=True was passed
+        mock_check.assert_called_once_with(force=True)
+
+    def test_release_notes_truncation(self, mock_api):
+        """Test that long release notes are truncated."""
+        long_notes = "A" * 1000  # 1000 character release notes
+        result_with_long_notes = MockUpdateCheckResult(
+            update_available=True,
+            current_version="1.7.5",
+            latest_version="1.8.0",
+            version_info=MockVersionInfo(
+                version="1.8.0",
+                release_url="https://example.com",
+                release_notes=long_notes,
+                published_at="2025-01-15T10:00:00Z",
+                is_prerelease=False
+            ),
+            from_cache=False,
+            error=None
+        )
+
+        with patch('whisperjav.version_checker.check_for_updates', return_value=result_with_long_notes):
+            with patch('whisperjav.version_checker.get_update_notification_level', return_value='minor'):
+                result = mock_api.check_for_updates(force=True)
+
+        # Release notes should be truncated to 500 chars
+        assert len(result["release_notes"]) <= 500
+
+
+class TestNotificationLevels:
+    """Tests for notification level classification."""
+
+    @pytest.mark.parametrize("current,latest,expected_level", [
+        ("1.7.5", "1.7.6", "patch"),      # Patch update
+        ("1.7.5", "1.8.0", "minor"),      # Minor update
+        ("1.7.5", "2.0.0", "major"),      # Major update
+        ("1.7.5", "1.7.5", "none"),       # No update
+    ])
+    def test_notification_level_classification(self, mock_api, current, latest, expected_level):
+        """Test that notification levels are correctly classified."""
+        result = MockUpdateCheckResult(
+            update_available=(current != latest),
+            current_version=current,
+            latest_version=latest,
+            version_info=MockVersionInfo(
+                version=latest,
+                release_url="https://example.com",
+                release_notes="Test notes",
+                published_at="2025-01-15T10:00:00Z",
+                is_prerelease=False
+            ) if current != latest else None,
+            from_cache=False,
+            error=None
+        )
+
+        with patch('whisperjav.version_checker.check_for_updates', return_value=result):
+            with patch('whisperjav.version_checker.get_update_notification_level', return_value=expected_level):
+                response = mock_api.check_for_updates(force=True)
+
+        if current != latest:
+            assert response.get("notification_level") == expected_level
+
+
+# =============================================================================
+# dismiss_update_notification() Tests
+# =============================================================================
+
+class TestDismissUpdateNotification:
+    """Tests for dismiss_update_notification() API method."""
+
+    def test_dismiss_creates_file(self, mock_api, tmp_path, monkeypatch):
+        """Test that dismissing creates a dismissal file."""
+        monkeypatch.setenv('LOCALAPPDATA', str(tmp_path))
+
+        result = mock_api.dismiss_update_notification("1.8.0")
+
+        assert result["success"] is True
+
+        # Check file was created
+        dismissal_file = tmp_path / '.whisperjav_cache' / 'update_dismissed.json'
+        assert dismissal_file.exists()
+
+        # Verify contents
+        with open(dismissal_file) as f:
+            data = json.load(f)
+        assert data["version"] == "1.8.0"
+        assert "dismissed_at" in data
+
+    def test_dismiss_overwrites_previous(self, mock_api, tmp_path, monkeypatch):
+        """Test that new dismissal overwrites previous."""
+        monkeypatch.setenv('LOCALAPPDATA', str(tmp_path))
+
+        # First dismissal
+        mock_api.dismiss_update_notification("1.7.6")
+
+        # Second dismissal
+        mock_api.dismiss_update_notification("1.8.0")
+
+        # Check only latest version is stored
+        dismissal_file = tmp_path / '.whisperjav_cache' / 'update_dismissed.json'
+        with open(dismissal_file) as f:
+            data = json.load(f)
+        assert data["version"] == "1.8.0"
+
+    def test_dismiss_error_handling(self, mock_api, monkeypatch):
+        """Test error handling when dismissal fails."""
+        # Mock the file open to raise an error
+        with patch('builtins.open', side_effect=PermissionError("Access denied")):
+            with patch('pathlib.Path.mkdir'):  # Don't fail on mkdir
+                result = mock_api.dismiss_update_notification("1.8.0")
+
+        # Should return error, not crash
+        assert result["success"] is False
+        assert "error" in result
+
+
+# =============================================================================
+# get_dismissed_update() Tests
+# =============================================================================
+
+class TestGetDismissedUpdate:
+    """Tests for get_dismissed_update() API method."""
+
+    def test_get_dismissed_when_exists(self, mock_api, tmp_path, monkeypatch):
+        """Test retrieving dismissed update info."""
+        monkeypatch.setenv('LOCALAPPDATA', str(tmp_path))
+
+        # Create dismissal first
+        mock_api.dismiss_update_notification("1.8.0")
+
+        result = mock_api.get_dismissed_update()
+
+        assert result["success"] is True
+        assert result["dismissed"]["version"] == "1.8.0"
+        assert "dismissed_at" in result["dismissed"]
+
+    def test_get_dismissed_when_not_exists(self, mock_api, tmp_path, monkeypatch):
+        """Test when no dismissal exists."""
+        monkeypatch.setenv('LOCALAPPDATA', str(tmp_path))
+        # Don't create any dismissal file
+
+        result = mock_api.get_dismissed_update()
+
+        assert result["success"] is True
+        assert result["dismissed"] is None
+
+
+# =============================================================================
+# open_url() Tests
+# =============================================================================
+
+class TestOpenUrl:
+    """Tests for open_url() API method."""
+
+    def test_open_url_success(self, mock_api):
+        """Test successful URL opening."""
+        with patch('webbrowser.open', return_value=True) as mock_open:
+            result = mock_api.open_url("https://github.com/meizhong986/WhisperJAV/releases")
+
+        assert result["success"] is True
+        mock_open.assert_called_once_with("https://github.com/meizhong986/WhisperJAV/releases")
+
+    def test_open_url_github_releases(self, mock_api):
+        """Test opening GitHub releases page specifically."""
+        with patch('webbrowser.open', return_value=True) as mock_open:
+            result = mock_api.open_url("https://github.com/meizhong986/WhisperJAV/releases")
+
+        assert result["success"] is True
+        # Verify the exact URL used for major updates
+        mock_open.assert_called_with("https://github.com/meizhong986/WhisperJAV/releases")
+
+    def test_open_url_error_handling(self, mock_api):
+        """Test error handling when browser fails to open."""
+        with patch('webbrowser.open', side_effect=Exception("No browser available")):
+            result = mock_api.open_url("https://example.com")
+
+        assert result["success"] is False
+        assert "No browser available" in result["error"]
+
+    def test_open_url_various_urls(self, mock_api):
+        """Test opening various URL types."""
+        urls = [
+            "https://github.com/meizhong986/WhisperJAV",
+            "https://github.com/meizhong986/WhisperJAV/releases/tag/v1.8.0",
+            "https://github.com/meizhong986/WhisperJAV/issues",
+        ]
+
+        for url in urls:
+            with patch('webbrowser.open', return_value=True) as mock_open:
+                result = mock_api.open_url(url)
+                assert result["success"] is True
+                mock_open.assert_called_with(url)
+
+
+# =============================================================================
+# start_update() Tests
+# =============================================================================
+
+class TestStartUpdate:
+    """Tests for start_update() API method."""
+
+    def test_start_update_spawns_wrapper(self, mock_api, monkeypatch):
+        """Test that start_update spawns the update wrapper process."""
+        mock_popen = MagicMock()
+
+        with patch('subprocess.Popen', mock_popen):
+            with patch.object(Path, 'exists', return_value=True):
+                result = mock_api.start_update(wheel_only=False)
+
+        assert result["success"] is True
+        assert result["should_exit"] is True
+        assert "Update process started" in result["message"]
+
+        # Verify Popen was called
+        mock_popen.assert_called_once()
+        call_args = mock_popen.call_args
+        cmd = call_args[0][0]
+
+        # Should include update_wrapper module
+        assert "-m" in cmd
+        assert "whisperjav.update_wrapper" in cmd
+
+    def test_start_update_with_wheel_only(self, mock_api):
+        """Test start_update with wheel_only=True."""
+        mock_popen = MagicMock()
+
+        with patch('subprocess.Popen', mock_popen):
+            with patch.object(Path, 'exists', return_value=True):
+                result = mock_api.start_update(wheel_only=True)
+
+        assert result["success"] is True
+
+        # Verify --wheel-only flag is present
+        call_args = mock_popen.call_args
+        cmd = call_args[0][0]
+        assert "--wheel-only" in cmd
+
+    def test_start_update_includes_pid(self, mock_api):
+        """Test that start_update passes current PID to wrapper."""
+        mock_popen = MagicMock()
+        current_pid = os.getpid()
+
+        with patch('subprocess.Popen', mock_popen):
+            with patch.object(Path, 'exists', return_value=True):
+                mock_api.start_update()
+
+        call_args = mock_popen.call_args
+        cmd = call_args[0][0]
+
+        # Should include --pid argument
+        assert "--pid" in cmd
+        pid_idx = cmd.index("--pid")
+        assert cmd[pid_idx + 1] == str(current_pid)
+
+    def test_start_update_error_handling(self, mock_api):
+        """Test error handling when spawn fails."""
+        with patch('subprocess.Popen', side_effect=Exception("Spawn failed")):
+            with patch.object(Path, 'exists', return_value=True):
+                result = mock_api.start_update()
+
+        assert result["success"] is False
+        assert "Spawn failed" in result["error"]
+        assert result["should_exit"] is False
+
+    @pytest.mark.skipif(sys.platform != "win32", reason="Windows-specific test")
+    def test_start_update_windows_detached(self, mock_api):
+        """Test that Windows uses DETACHED_PROCESS flag."""
+        mock_popen = MagicMock()
+
+        with patch('subprocess.Popen', mock_popen):
+            with patch.object(Path, 'exists', return_value=True):
+                mock_api.start_update()
+
+        call_kwargs = mock_popen.call_args[1]
+
+        # Should use Windows-specific creation flags
+        assert "creationflags" in call_kwargs
+        expected_flags = subprocess.DETACHED_PROCESS | subprocess.CREATE_NEW_PROCESS_GROUP
+        assert call_kwargs["creationflags"] == expected_flags
+
+
+# =============================================================================
+# Integration Tests
+# =============================================================================
+
+class TestUpdateCheckIntegration:
+    """Integration tests for the complete update check flow."""
+
+    def test_full_update_check_flow(self, mock_api, sample_update_result, tmp_path, monkeypatch):
+        """Test complete flow: check -> show update -> user action."""
+        monkeypatch.setenv('LOCALAPPDATA', str(tmp_path))
+
+        # Step 1: Check for updates
+        with patch('whisperjav.version_checker.check_for_updates', return_value=sample_update_result):
+            with patch('whisperjav.version_checker.get_update_notification_level', return_value='minor'):
+                check_result = mock_api.check_for_updates(force=True)
+
+        assert check_result["update_available"] is True
+        assert check_result["notification_level"] == "minor"
+
+        # Step 2: User decides to dismiss
+        dismiss_result = mock_api.dismiss_update_notification(check_result["latest_version"])
+        assert dismiss_result["success"] is True
+
+        # Step 3: Verify dismissal was recorded
+        dismissed = mock_api.get_dismissed_update()
+        assert dismissed["dismissed"]["version"] == "1.8.0"
+
+    def test_major_update_triggers_download(self, mock_api, tmp_path, monkeypatch):
+        """Test that major updates trigger download page instead of update."""
+        monkeypatch.setenv('LOCALAPPDATA', str(tmp_path))
+
+        major_update = MockUpdateCheckResult(
+            update_available=True,
+            current_version="1.7.5",
+            latest_version="2.0.0",
+            version_info=MockVersionInfo(
+                version="2.0.0",
+                release_url="https://github.com/meizhong986/whisperjav/releases/tag/v2.0.0",
+                release_notes="## Breaking Changes\n- New architecture",
+                published_at="2025-01-15T10:00:00Z",
+                is_prerelease=False
+            ),
+            from_cache=False,
+            error=None
+        )
+
+        with patch('whisperjav.version_checker.check_for_updates', return_value=major_update):
+            with patch('whisperjav.version_checker.get_update_notification_level', return_value='major'):
+                result = mock_api.check_for_updates(force=True)
+
+        # For major updates, frontend should use open_url() instead of start_update()
+        assert result["notification_level"] == "major"
+
+        # Verify open_url works for this case
+        with patch('webbrowser.open', return_value=True) as mock_open:
+            url_result = mock_api.open_url("https://github.com/meizhong986/WhisperJAV/releases")
+
+        assert url_result["success"] is True
+
+    def test_patch_update_triggers_inline_update(self, mock_api, tmp_path, monkeypatch):
+        """Test that patch updates trigger inline update."""
+        monkeypatch.setenv('LOCALAPPDATA', str(tmp_path))
+
+        patch_update = MockUpdateCheckResult(
+            update_available=True,
+            current_version="1.7.5",
+            latest_version="1.7.6",
+            version_info=MockVersionInfo(
+                version="1.7.6",
+                release_url="https://github.com/meizhong986/whisperjav/releases/tag/v1.7.6",
+                release_notes="## Bug Fixes\n- Fixed MPS fallback",
+                published_at="2025-01-15T10:00:00Z",
+                is_prerelease=False
+            ),
+            from_cache=False,
+            error=None
+        )
+
+        with patch('whisperjav.version_checker.check_for_updates', return_value=patch_update):
+            with patch('whisperjav.version_checker.get_update_notification_level', return_value='patch'):
+                result = mock_api.check_for_updates(force=True)
+
+        # For patch updates, frontend should use start_update()
+        assert result["notification_level"] == "patch"
+
+        # Verify start_update works
+        mock_popen = MagicMock()
+        with patch('subprocess.Popen', mock_popen):
+            with patch.object(Path, 'exists', return_value=True):
+                update_result = mock_api.start_update(wheel_only=False)
+
+        assert update_result["success"] is True
+        assert update_result["should_exit"] is True
+
+
+# =============================================================================
+# Edge Cases and Error Handling
+# =============================================================================
+
+class TestEdgeCases:
+    """Tests for edge cases and error conditions."""
+
+    def test_network_timeout_handling(self, mock_api):
+        """Test handling of network timeout."""
+        timeout_result = MockUpdateCheckResult(
+            update_available=False,
+            current_version="1.7.5",
+            latest_version=None,
+            version_info=None,
+            from_cache=False,
+            error="Connection timed out after 10 seconds"
+        )
+
+        with patch('whisperjav.version_checker.check_for_updates', return_value=timeout_result):
+            result = mock_api.check_for_updates(force=True)
+
+        assert result["update_available"] is False
+        assert "timed out" in result["error"]
+
+    def test_invalid_version_format(self, mock_api):
+        """Test handling of invalid version format."""
+        invalid_result = MockUpdateCheckResult(
+            update_available=False,
+            current_version="unknown",
+            latest_version=None,
+            version_info=None,
+            from_cache=False,
+            error="Could not parse version"
+        )
+
+        with patch('whisperjav.version_checker.check_for_updates', return_value=invalid_result):
+            result = mock_api.check_for_updates(force=True)
+
+        assert result["success"] is True  # API call succeeded
+        assert "error" in result
+
+    def test_prerelease_handling(self, mock_api):
+        """Test that prerelease versions are identified."""
+        prerelease = MockUpdateCheckResult(
+            update_available=True,
+            current_version="1.7.5",
+            latest_version="1.8.0-rc1",
+            version_info=MockVersionInfo(
+                version="1.8.0-rc1",
+                release_url="https://example.com",
+                release_notes="Release candidate",
+                published_at="2025-01-15T10:00:00Z",
+                is_prerelease=True
+            ),
+            from_cache=False,
+            error=None
+        )
+
+        with patch('whisperjav.version_checker.check_for_updates', return_value=prerelease):
+            with patch('whisperjav.version_checker.get_update_notification_level', return_value='minor'):
+                result = mock_api.check_for_updates(force=True)
+
+        assert result["is_prerelease"] is True
+
+    def test_empty_release_notes(self, mock_api):
+        """Test handling of empty release notes."""
+        empty_notes = MockUpdateCheckResult(
+            update_available=True,
+            current_version="1.7.5",
+            latest_version="1.7.6",
+            version_info=MockVersionInfo(
+                version="1.7.6",
+                release_url="https://example.com",
+                release_notes="",
+                published_at="2025-01-15T10:00:00Z",
+                is_prerelease=False
+            ),
+            from_cache=False,
+            error=None
+        )
+
+        with patch('whisperjav.version_checker.check_for_updates', return_value=empty_notes):
+            with patch('whisperjav.version_checker.get_update_notification_level', return_value='patch'):
+                result = mock_api.check_for_updates(force=True)
+
+        assert result["release_notes"] == ""
+
+
+if __name__ == "__main__":
+    pytest.main([__file__, "-v"])

tests/test_update_check_frontend.py

@@ -0,0 +1,517 @@
+#!/usr/bin/env python3
+"""
+Frontend validation tests for WhisperJAV GUI Update Check feature.
+
+Tests the HTML structure, CSS selectors, and JavaScript syntax for the
+"Check for Updates" modal and menu integration.
+
+These tests validate:
+- Required HTML elements exist with correct IDs
+- CSS classes are properly defined
+- JavaScript UpdateCheckManager has required methods
+- Event handlers reference existing elements
+
+Run with: pytest tests/test_update_check_frontend.py -v
+"""
+
+import os
+import re
+import pytest
+from pathlib import Path
+
+
+# =============================================================================
+# Fixtures
+# =============================================================================
+
+@pytest.fixture
+def assets_dir():
+    """Get the path to the webview_gui assets directory."""
+    repo_root = Path(__file__).parent.parent
+    return repo_root / "whisperjav" / "webview_gui" / "assets"
+
+
+@pytest.fixture
+def index_html(assets_dir):
+    """Load the index.html content."""
+    html_path = assets_dir / "index.html"
+    assert html_path.exists(), f"index.html not found at {html_path}"
+    return html_path.read_text(encoding="utf-8")
+
+
+@pytest.fixture
+def app_js(assets_dir):
+    """Load the app.js content."""
+    js_path = assets_dir / "app.js"
+    assert js_path.exists(), f"app.js not found at {js_path}"
+    return js_path.read_text(encoding="utf-8")
+
+
+@pytest.fixture
+def style_css(assets_dir):
+    """Load the style.css content."""
+    css_path = assets_dir / "style.css"
+    assert css_path.exists(), f"style.css not found at {css_path}"
+    return css_path.read_text(encoding="utf-8")
+
+
+# =============================================================================
+# HTML Structure Tests
+# =============================================================================
+
+class TestHtmlStructure:
+    """Tests for required HTML elements in index.html."""
+
+    def test_theme_menu_exists(self, index_html):
+        """Test that theme menu container exists."""
+        assert 'id="themeMenu"' in index_html
+
+    def test_check_updates_button_exists(self, index_html):
+        """Test that Check for Updates button exists."""
+        assert 'id="checkUpdatesBtn"' in index_html
+
+    def test_check_updates_button_in_theme_menu(self, index_html):
+        """Test that button is inside theme menu."""
+        # Extract theme menu content
+        menu_match = re.search(r'id="themeMenu"[^>]*>(.*?)</div>\s*</div>', index_html, re.DOTALL)
+        assert menu_match, "Could not find themeMenu content"
+        menu_content = menu_match.group(1)
+        assert 'checkUpdatesBtn' in menu_content
+
+    def test_theme_menu_divider_exists(self, index_html):
+        """Test that menu divider exists before update check option."""
+        assert 'class="theme-menu-divider"' in index_html
+
+    def test_update_check_icon_exists(self, index_html):
+        """Test that update icon span exists."""
+        assert 'class="update-check-icon"' in index_html
+
+    def test_modal_overlay_exists(self, index_html):
+        """Test that update check modal overlay exists."""
+        assert 'id="updateCheckModal"' in index_html
+
+    def test_modal_close_button_exists(self, index_html):
+        """Test that modal close button exists."""
+        assert 'id="updateCheckModalClose"' in index_html
+
+    def test_loading_state_exists(self, index_html):
+        """Test that loading state container exists."""
+        assert 'id="updateCheckLoading"' in index_html
+
+    def test_result_state_exists(self, index_html):
+        """Test that result state container exists."""
+        assert 'id="updateCheckResult"' in index_html
+
+    def test_error_state_exists(self, index_html):
+        """Test that error state container exists."""
+        assert 'id="updateCheckError"' in index_html
+
+    def test_version_elements_exist(self, index_html):
+        """Test that version display elements exist."""
+        required_ids = [
+            "updateCurrentVersion",
+            "updateLatestVersion",
+            "updateTypeBadge",
+        ]
+        for element_id in required_ids:
+            assert f'id="{element_id}"' in index_html, f"Missing element: {element_id}"
+
+    def test_up_to_date_message_exists(self, index_html):
+        """Test that up-to-date message exists."""
+        assert 'id="updateUpToDate"' in index_html
+
+    def test_update_available_section_exists(self, index_html):
+        """Test that update available section exists."""
+        assert 'id="updateAvailableSection"' in index_html
+
+    def test_release_notes_container_exists(self, index_html):
+        """Test that release notes container exists."""
+        assert 'id="updateReleaseNotes"' in index_html
+
+    def test_action_buttons_exist(self, index_html):
+        """Test that action buttons exist."""
+        required_buttons = [
+            "updateCheckLater",
+            "updateCheckNow",
+            "updateCheckDownload",
+        ]
+        for button_id in required_buttons:
+            assert f'id="{button_id}"' in index_html, f"Missing button: {button_id}"
+
+    def test_modal_has_correct_class(self, index_html):
+        """Test that modal uses correct overlay class."""
+        assert 'class="modal-overlay" id="updateCheckModal"' in index_html or \
+               'id="updateCheckModal" class="modal-overlay"' in index_html
+
+
+# =============================================================================
+# CSS Validation Tests
+# =============================================================================
+
+class TestCssClasses:
+    """Tests for required CSS classes in style.css."""
+
+    def test_theme_menu_divider_style(self, style_css):
+        """Test that theme-menu-divider class is defined."""
+        assert ".theme-menu-divider" in style_css
+
+    def test_update_check_option_style(self, style_css):
+        """Test that update-check-option class is defined."""
+        assert ".update-check-option" in style_css
+
+    def test_update_check_icon_style(self, style_css):
+        """Test that update-check-icon class is defined."""
+        assert ".update-check-icon" in style_css
+
+    def test_update_check_loading_style(self, style_css):
+        """Test that update-check-loading class is defined."""
+        assert ".update-check-loading" in style_css
+
+    def test_spinner_style(self, style_css):
+        """Test that spinner class is defined."""
+        assert ".spinner" in style_css or ".update-check-loading .spinner" in style_css
+
+    def test_version_info_styles(self, style_css):
+        """Test that version info styles are defined."""
+        assert ".version-info" in style_css
+        assert ".version-row" in style_css
+
+    def test_update_badge_styles(self, style_css):
+        """Test that update badge styles are defined."""
+        assert ".update-badge" in style_css
+        assert ".update-badge.patch" in style_css
+        assert ".update-badge.minor" in style_css
+        assert ".update-badge.major" in style_css
+        assert ".update-badge.critical" in style_css
+
+    def test_up_to_date_style(self, style_css):
+        """Test that up-to-date message style is defined."""
+        assert ".update-up-to-date" in style_css
+
+    def test_release_notes_styles(self, style_css):
+        """Test that release notes styles are defined."""
+        assert ".release-notes-details" in style_css
+        assert ".release-notes-content" in style_css
+
+    def test_update_check_error_style(self, style_css):
+        """Test that error state style is defined."""
+        assert ".update-check-error" in style_css
+
+    def test_spin_animation_exists(self, style_css):
+        """Test that spin keyframe animation exists."""
+        assert "@keyframes spin" in style_css
+
+
+class TestCssColorScheme:
+    """Tests for consistent color scheme in badge styles."""
+
+    def test_patch_badge_color(self, style_css):
+        """Test that patch badge has green color."""
+        # Find the patch badge rule
+        match = re.search(r'\.update-badge\.patch\s*\{([^}]+)\}', style_css)
+        assert match, "Patch badge style not found"
+        assert "#28a745" in match.group(1) or "green" in match.group(1).lower()
+
+    def test_minor_badge_color(self, style_css):
+        """Test that minor badge has blue color."""
+        match = re.search(r'\.update-badge\.minor\s*\{([^}]+)\}', style_css)
+        assert match, "Minor badge style not found"
+        assert "#17a2b8" in match.group(1) or "blue" in match.group(1).lower()
+
+    def test_major_badge_color(self, style_css):
+        """Test that major badge has orange color."""
+        match = re.search(r'\.update-badge\.major\s*\{([^}]+)\}', style_css)
+        assert match, "Major badge style not found"
+        assert "#fd7e14" in match.group(1) or "orange" in match.group(1).lower()
+
+    def test_critical_badge_color(self, style_css):
+        """Test that critical badge has red color."""
+        match = re.search(r'\.update-badge\.critical\s*\{([^}]+)\}', style_css)
+        assert match, "Critical badge style not found"
+        assert "#dc3545" in match.group(1) or "red" in match.group(1).lower()
+
+
+# =============================================================================
+# JavaScript Structure Tests
+# =============================================================================
+
+class TestJavaScriptStructure:
+    """Tests for JavaScript UpdateCheckManager in app.js."""
+
+    def test_update_check_manager_exists(self, app_js):
+        """Test that UpdateCheckManager object is defined."""
+        assert "const UpdateCheckManager" in app_js or "var UpdateCheckManager" in app_js
+
+    def test_init_method_exists(self, app_js):
+        """Test that init() method exists."""
+        # Find UpdateCheckManager definition
+        manager_match = re.search(
+            r'const UpdateCheckManager\s*=\s*\{(.*?)\n\};',
+            app_js,
+            re.DOTALL
+        )
+        assert manager_match, "UpdateCheckManager not found"
+        manager_body = manager_match.group(1)
+        assert "init()" in manager_body or "init:" in manager_body
+
+    def test_show_method_exists(self, app_js):
+        """Test that show() method exists."""
+        assert re.search(r'(async\s+)?show\s*\(\s*\)', app_js)
+
+    def test_close_method_exists(self, app_js):
+        """Test that close() method exists."""
+        assert "close()" in app_js or "close:" in app_js
+
+    def test_show_loading_method_exists(self, app_js):
+        """Test that showLoading() method exists."""
+        assert "showLoading()" in app_js or "showLoading:" in app_js
+
+    def test_show_result_method_exists(self, app_js):
+        """Test that showResult() method exists."""
+        assert "showResult(" in app_js
+
+    def test_show_error_method_exists(self, app_js):
+        """Test that showError() method exists."""
+        assert "showError(" in app_js
+
+    def test_start_update_method_exists(self, app_js):
+        """Test that startUpdate() method exists."""
+        assert "startUpdate()" in app_js or "startUpdate:" in app_js
+
+    def test_open_download_page_method_exists(self, app_js):
+        """Test that openDownloadPage() method exists."""
+        assert "openDownloadPage()" in app_js or "openDownloadPage:" in app_js
+
+    def test_parse_markdown_method_exists(self, app_js):
+        """Test that parseMarkdown() method exists."""
+        assert "parseMarkdown(" in app_js
+
+    def test_init_called_on_dom_ready(self, app_js):
+        """Test that UpdateCheckManager.init() is called on DOMContentLoaded."""
+        # Find the DOMContentLoaded handler
+        dom_ready_match = re.search(
+            r"DOMContentLoaded.*?\{(.*?)\}\s*\)",
+            app_js,
+            re.DOTALL
+        )
+        assert dom_ready_match, "DOMContentLoaded handler not found"
+        handler_body = dom_ready_match.group(1)
+        assert "UpdateCheckManager.init()" in handler_body
+
+
+class TestJavaScriptApiCalls:
+    """Tests for correct API call usage in JavaScript."""
+
+    def test_check_for_updates_api_call(self, app_js):
+        """Test that check_for_updates API is called correctly."""
+        # Should call with force=true for modal
+        assert "pywebview.api.check_for_updates" in app_js
+        assert "check_for_updates(true)" in app_js
+
+    def test_open_url_api_call(self, app_js):
+        """Test that open_url API is called for download page."""
+        assert "pywebview.api.open_url" in app_js
+        # Should include GitHub releases URL
+        assert "github.com/meizhong986/WhisperJAV/releases" in app_js
+
+    def test_start_update_delegation(self, app_js):
+        """Test that startUpdate delegates to UpdateManager."""
+        # Should call UpdateManager.startUpdate() not api.start_update directly
+        assert "UpdateManager.startUpdate()" in app_js
+
+
+class TestJavaScriptEventHandlers:
+    """Tests for event handler setup in JavaScript."""
+
+    def test_menu_button_click_handler(self, app_js):
+        """Test that menu button has click handler."""
+        assert "checkUpdatesBtn" in app_js
+        assert "addEventListener" in app_js
+
+    def test_modal_close_button_handler(self, app_js):
+        """Test that modal close button has handler."""
+        assert "updateCheckModalClose" in app_js
+
+    def test_later_button_handler(self, app_js):
+        """Test that Later button has handler."""
+        assert "updateCheckLater" in app_js
+
+    def test_update_now_button_handler(self, app_js):
+        """Test that Update Now button has handler."""
+        assert "updateCheckNow" in app_js
+
+    def test_download_button_handler(self, app_js):
+        """Test that Download button has handler."""
+        assert "updateCheckDownload" in app_js
+
+    def test_escape_key_handler(self, app_js):
+        """Test that Escape key closes modal."""
+        assert "Escape" in app_js
+
+    def test_overlay_click_closes_modal(self, app_js):
+        """Test that clicking overlay closes modal."""
+        # Should check if click target is the modal itself
+        assert "e.target === this.modal" in app_js or \
+               "event.target === this.modal" in app_js
+
+
+class TestJavaScriptLogic:
+    """Tests for JavaScript business logic."""
+
+    def test_major_update_shows_download_button(self, app_js):
+        """Test that major updates show Download button."""
+        # Check for level === 'major' condition
+        assert "'major'" in app_js
+        assert "updateCheckDownload" in app_js
+
+    def test_patch_minor_show_update_button(self, app_js):
+        """Test that patch/minor updates show Update Now button."""
+        assert "updateCheckNow" in app_js
+
+    def test_notification_level_badge_update(self, app_js):
+        """Test that badge class is updated based on level."""
+        assert "update-badge" in app_js
+        assert "className" in app_js or "classList" in app_js
+
+    def test_version_display_formatting(self, app_js):
+        """Test that versions are prefixed with 'v'."""
+        assert "'v' + result.current_version" in app_js or \
+               '"v" + result.current_version' in app_js
+
+    def test_up_to_date_display_logic(self, app_js):
+        """Test that up-to-date message shows when no update."""
+        assert "updateUpToDate" in app_js
+        # Should show when update_available is false
+        assert "update_available" in app_js.lower() or "updateavailable" in app_js.lower()
+
+
+# =============================================================================
+# Theme Compatibility Tests
+# =============================================================================
+
+class TestThemeCompatibility:
+    """Tests for theme file compatibility."""
+
+    @pytest.fixture
+    def theme_files(self, assets_dir):
+        """Get all theme CSS files."""
+        return list(assets_dir.glob("style.*.css"))
+
+    def test_themes_import_base_styles(self, assets_dir):
+        """Test that theme files import base style.css."""
+        theme_files = [
+            "style.google.css",
+            "style.carbon.css",
+            "style.primer.css",
+        ]
+        for theme_file in theme_files:
+            theme_path = assets_dir / theme_file
+            if theme_path.exists():
+                content = theme_path.read_text(encoding="utf-8")
+                assert '@import url("style.css")' in content or \
+                       "@import url('style.css')" in content, \
+                       f"{theme_file} should import base styles"
+
+    def test_css_uses_variables(self, style_css):
+        """Test that update check styles use CSS variables where appropriate."""
+        # Key elements should use CSS variables for theme compatibility
+        update_section_match = re.search(
+            r'/\* Update Check Modal \*/.*?(?=/\*|$)',
+            style_css,
+            re.DOTALL
+        )
+        if update_section_match:
+            update_css = update_section_match.group(0)
+            # Should use variables like var(--border-color)
+            assert "var(--" in update_css
+
+
+# =============================================================================
+# Accessibility Tests
+# =============================================================================
+
+class TestAccessibility:
+    """Tests for accessibility features."""
+
+    def test_modal_close_has_aria_label(self, index_html):
+        """Test that close button has aria-label."""
+        close_match = re.search(r'id="updateCheckModalClose"[^>]*>', index_html)
+        assert close_match, "Close button not found"
+        close_tag = close_match.group(0)
+        assert 'aria-label' in close_tag
+
+    def test_menu_items_have_role(self, index_html):
+        """Test that menu items have proper role."""
+        # Check updates button should have role="menuitem"
+        assert 'id="checkUpdatesBtn"' in index_html
+        btn_match = re.search(r'id="checkUpdatesBtn"[^>]*>', index_html)
+        if btn_match:
+            btn_tag = btn_match.group(0)
+            assert 'role="menuitem"' in btn_tag
+
+    def test_buttons_have_labels(self, index_html):
+        """Test that buttons have visible text or aria-label."""
+        # All action buttons should have text content
+        assert "Update Now" in index_html
+        assert "Download" in index_html
+        assert "Close" in index_html or "Later" in index_html
+
+
+# =============================================================================
+# Consistency Tests
+# =============================================================================
+
+class TestConsistency:
+    """Tests for consistency between HTML, CSS, and JS."""
+
+    def test_html_ids_match_js_selectors(self, index_html, app_js):
+        """Test that HTML element IDs match JavaScript selectors."""
+        html_ids = [
+            "updateCheckModal",
+            "updateCheckModalClose",
+            "checkUpdatesBtn",
+            "updateCheckLoading",
+            "updateCheckResult",
+            "updateCheckError",
+            "updateCurrentVersion",
+            "updateLatestVersion",
+            "updateTypeBadge",
+            "updateUpToDate",
+            "updateAvailableSection",
+            "updateReleaseNotes",
+            "updateCheckLater",
+            "updateCheckNow",
+            "updateCheckDownload",
+        ]
+
+        for html_id in html_ids:
+            assert f'id="{html_id}"' in index_html, f"HTML missing id: {html_id}"
+            # JS should reference this ID
+            assert html_id in app_js, f"JS not referencing id: {html_id}"
+
+    def test_css_classes_match_html(self, index_html, style_css):
+        """Test that CSS classes used in HTML are defined."""
+        html_classes = [
+            "theme-menu-divider",
+            "update-check-option",
+            "update-check-icon",
+            "update-check-loading",
+            "update-check-result",
+            "update-check-error",
+            "version-info",
+            "version-row",
+            "update-badge",
+            "update-up-to-date",
+            "release-notes-details",
+            "release-notes-content",
+        ]
+
+        for css_class in html_classes:
+            # Check if class is used in HTML (may be part of multi-class declaration)
+            assert css_class in index_html, f"HTML should use class: {css_class}"
+            # Check if class is defined in CSS
+            assert f".{css_class}" in style_css, f"CSS should define class: {css_class}"
+
+
+if __name__ == "__main__":
+    pytest.main([__file__, "-v"])

whisperjav/__version__.py

@@ -2,16 +2,16 @@
 """Version information for WhisperJAV."""
 
 # PEP 440 compliant version for pip/wheel
-__version__ = "1.7.4"
+__version__ = "1.7.5"
 
 # Human-readable version for display in UI
-__version_display__ = "1.7.4"
+__version_display__ = "1.7.5"
 
 # Version metadata
 __version_info__ = {
     "major": 1,
     "minor": 7,
-    "patch": 4,
+    "patch": 5,
     "release": "",
     "architecture": "v4.4"
 }

whisperjav/config/components/base.py

@@ -102,7 +102,13 @@ class ASRComponent(ComponentBase):
     supported_tasks: List[str] = ["transcribe"]  # "transcribe", "translate"
     compatible_vad: List[str] = []  # List of compatible VAD component names
 
-    # === Compute ===
+    # === Compute (Preferred Defaults) ===
+    # NOTE: These are "preferred" defaults for optimal performance.
+    # The resolver applies runtime device detection via get_best_device()
+    # and auto-selects compute_type based on device and provider:
+    #   - CUDA → int8_float16 for CTranslate2, float16 for PyTorch
+    #   - CPU  → int8 for CTranslate2, float32 for PyTorch
+    #   - MPS  → float16 for PyTorch, falls back to CPU+int8 for CTranslate2
     default_device: str = "cuda"
     default_compute_type: str = "float16"
 

whisperjav/config/legacy.py

@@ -135,6 +135,8 @@ def resolve_legacy_pipeline(
     sensitivity: str = "balanced",
     task: str = "transcribe",
     overrides: Optional[Dict[str, Any]] = None,
+    device: Optional[str] = None,
+    compute_type: Optional[str] = None,
 ) -> Dict[str, Any]:
     """
     Resolve configuration from legacy pipeline name.
@@ -146,12 +148,16 @@ def resolve_legacy_pipeline(
         sensitivity: Sensitivity level
         task: Task type
         overrides: Parameter overrides
+        device: Device override (None/'auto' = auto-detect, 'cuda'/'cpu' = explicit)
+        compute_type: Compute type override (None/'auto' = provider-specific default)
 
     Returns:
         Resolved configuration dictionary.
 
     Example:
         >>> config = resolve_legacy_pipeline('balanced', 'aggressive')
+        >>> # With explicit hardware:
+        >>> config = resolve_legacy_pipeline('balanced', device='cuda', compute_type='int8_float16')
     """
     if pipeline_name not in LEGACY_PIPELINES:
         available = list(LEGACY_PIPELINES.keys())
@@ -167,6 +173,8 @@ def resolve_legacy_pipeline(
         task=task,
         features=pipeline_def["features"],
         overrides=overrides,
+        device=device,
+        compute_type=compute_type,
     )
 
     # Add legacy compatibility fields
@@ -346,6 +354,8 @@ def resolve_ensemble_config(
     task: str = "transcribe",
     features: Optional[List[str]] = None,
     overrides: Optional[Dict[str, Any]] = None,
+    device: Optional[str] = None,
+    compute_type: Optional[str] = None,
 ) -> Dict[str, Any]:
     """
     Resolve configuration for ensemble mode (direct component specification).
@@ -360,6 +370,8 @@ def resolve_ensemble_config(
         features: List of feature names (e.g., ['auditok_scene_detection'])
         overrides: Parameter overrides in flat dot notation
                    (e.g., {'asr.beam_size': 10, 'vad.threshold': 0.25})
+        device: Device override (None/'auto' = auto-detect, 'cuda'/'cpu' = explicit)
+        compute_type: Compute type override (None/'auto' = provider-specific default)
 
     Returns:
         Resolved configuration in legacy structure for pipeline compatibility.
@@ -395,6 +407,8 @@ def resolve_ensemble_config(
         task=task,
         features=features or [],
         overrides=nested_overrides,
+        device=device,
+        compute_type=compute_type,
     )
 
     # Add ensemble-specific metadata

whisperjav/config/resolver_v3.py

@@ -23,9 +23,46 @@ from .components import (
     get_vad_registry,
     get_feature_registry,
 )
+from whisperjav.utils.device_detector import get_best_device
 
 logger = logging.getLogger("whisperjav")
 
+# CTranslate2-based providers support "auto" compute_type selection
+CTRANSLATE2_PROVIDERS = {"faster_whisper", "kotoba_faster_whisper"}
+
+
+def _get_compute_type_for_device(device: str, provider: str) -> str:
+    """
+    Select optimal compute_type based on device and provider.
+
+    CTranslate2 providers (faster_whisper, kotoba_faster_whisper):
+    - Returns "auto" to delegate selection to CTranslate2's internal logic
+    - CTranslate2 auto-selects based on GPU capability and CPU instruction sets
+    - Benefits: Automatic support for new GPUs (e.g., RTX 50XX Blackwell sm120)
+    - See: https://github.com/OpenNMT/CTranslate2/issues/1865
+
+    PyTorch providers (openai_whisper, stable_ts):
+    - CUDA/MPS: float16
+    - CPU: float32
+
+    Args:
+        device: Detected device ("cuda", "mps", "cpu")
+        provider: ASR provider name
+
+    Returns:
+        Optimal compute_type for the device/provider combination
+    """
+    if provider in CTRANSLATE2_PROVIDERS:
+        # Delegate to CTranslate2's internal "auto" selection logic
+        # This automatically handles:
+        # - GPU capability detection (int8_float16 for RTX 20/30/40, float16 for RTX 50)
+        # - CPU instruction sets (AVX2, AVX512)
+        # - Blackwell sm120 workaround (PR #1937)
+        return "auto"
+    else:
+        # PyTorch-based providers can use float16 on GPU/MPS, float32 on CPU
+        return "float16" if device in ("cuda", "mps") else "float32"
+
 
 def resolve_config_v3(
     asr: str,
@@ -34,6 +71,8 @@ def resolve_config_v3(
     task: str = "transcribe",
     features: Optional[List[str]] = None,
     overrides: Optional[Dict[str, Any]] = None,
+    device: Optional[str] = None,
+    compute_type: Optional[str] = None,
 ) -> Dict[str, Any]:
     """
     Resolve configuration from component selections.
@@ -45,6 +84,9 @@ def resolve_config_v3(
         task: Task type ('transcribe', 'translate')
         features: List of feature names (e.g., ['auditok_scene_detection'])
         overrides: Parameter overrides (e.g., {'asr.beam_size': 10})
+        device: Device override (None/'auto' = auto-detect, 'cuda'/'cpu' = explicit)
+        compute_type: Compute type override (None/'auto' = provider-specific default,
+                      'float16'/'float32'/'int8'/'int8_float16'/'int8_float32' = explicit)
 
     Returns:
         Resolved configuration dictionary.
@@ -58,6 +100,12 @@ def resolve_config_v3(
         ...     features=['auditok_scene_detection'],
         ...     overrides={'asr.beam_size': 10}
         ... )
+        >>> # With explicit hardware configuration:
+        >>> config = resolve_config_v3(
+        ...     asr='faster_whisper',
+        ...     device='cuda',
+        ...     compute_type='int8_float16'
+        ... )
     """
     if features is None:
         features = []
@@ -125,12 +173,38 @@ def resolve_config_v3(
     for feature_type, feature_config in feature_configs.items():
         _apply_overrides(feature_config, overrides, feature_type)
 
-    # Build model configuration
+    # Build model configuration with device/compute_type selection
+    # Priority: 1) User-provided value, 2) Auto-detection
+    #
+    # Device selection:
+    # - None or "auto" → auto-detect using get_best_device()
+    # - "cuda"/"cpu" → use explicit value (MPS fallback handled by ASR modules)
+    #
+    # Compute type selection:
+    # - None or "auto" → provider-specific default (CTranslate2 "auto" or PyTorch float16/32)
+    # - Explicit value → pass through to ASR module
+    if device and device != "auto":
+        selected_device = device
+        logger.debug(f"v3.0 resolver: Using user-specified device='{selected_device}'")
+    else:
+        selected_device = get_best_device()
+        logger.debug(f"v3.0 resolver: Auto-detected device='{selected_device}'")
+
+    if compute_type and compute_type != "auto":
+        selected_compute_type = compute_type
+        logger.debug(f"v3.0 resolver: Using user-specified compute_type='{selected_compute_type}'")
+    else:
+        selected_compute_type = _get_compute_type_for_device(selected_device, asr_component.provider)
+        logger.debug(
+            f"v3.0 resolver: Selected compute_type='{selected_compute_type}' "
+            f"for provider='{asr_component.provider}'"
+        )
+
     model_config = {
         'provider': asr_component.provider,
         'model_name': asr_component.model_id,
-        'device': asr_component.default_device,
-        'compute_type': asr_component.default_compute_type,
+        'device': selected_device,
+        'compute_type': selected_compute_type,
         'supported_tasks': asr_component.supported_tasks,
     }
 

whisperjav/ensemble/pass_worker.py

@@ -696,6 +696,11 @@ def _build_pipeline(
             logger.debug("Pass %s: Override hf_scene = %s", pass_number, pass_config["scene_detector"])
         elif pass_config.get("scene_detector") == "none":
             hf_defaults["hf_scene"] = "none"
+        # Apply source language override for transformers pipeline (fixes Issue #104)
+        # The language code (e.g., 'en', 'ja') is set in main.py from --language argument
+        if pass_config.get("language"):
+            hf_defaults["hf_language"] = pass_config["language"]
+            logger.debug("Pass %s: Override hf_language = %s", pass_number, pass_config["language"])
         # Apply speech enhancer override for transformers pipeline
         if pass_config.get("speech_enhancer"):
             enhancer_backend, enhancer_model = _parse_speech_enhancer(pass_config["speech_enhancer"])
@@ -737,8 +742,19 @@ def _build_pipeline(
         sensitivity=pass_config.get("sensitivity", "balanced"),
         task=asr_task,  # Derived from subs_language above
         overrides=pass_config.get("overrides"),
+        device=pass_config.get("device"),  # None = auto-detect
+        compute_type=pass_config.get("compute_type"),  # None = auto
     )
 
+    # Apply source language from pass_config (fixes Issue #104)
+    # The language code (e.g., 'en', 'ja') is set in main.py from --language argument
+    if pass_config.get("language"):
+        resolved_config["language"] = pass_config["language"]
+        # Also update decoder params so ASR uses correct language
+        if "params" in resolved_config and "decoder" in resolved_config["params"]:
+            resolved_config["params"]["decoder"]["language"] = pass_config["language"]
+        logger.debug("Pass %s: Applied source language = %s", pass_number, pass_config["language"])
+
     # Apply GUI-specified overrides for legacy pipelines
     _apply_gui_overrides(resolved_config, pass_config, pass_number)
 

whisperjav/main.py

@@ -211,6 +211,19 @@ def parse_arguments():
     parser.add_argument("--accept-cpu-mode", action="store_true",
                        help="Accept CPU-only mode without GPU warning (skip GPU performance check)")
 
+    # Hardware configuration (device and compute type override)
+    hardware_group = parser.add_argument_group("Hardware Configuration")
+    hardware_group.add_argument("--device", type=str, default=None,
+                               choices=["auto", "cuda", "cpu"],
+                               help="Device to use for ASR processing (default: auto-detect). "
+                                    "Note: MPS (Apple Silicon) is auto-detected but CTranslate2 backends "
+                                    "will fall back to CPU with Accelerate optimization.")
+    hardware_group.add_argument("--compute-type", type=str, default=None,
+                               choices=["auto", "float16", "float32", "int8", "int8_float16", "int8_float32"],
+                               help="Compute type for ASR processing (default: auto). "
+                                    "CTranslate2 backends (faster_whisper, kotoba) use 'auto' to optimize "
+                                    "based on GPU capability. PyTorch backends use float16 (GPU) or float32 (CPU).")
+
     # Path and logging
     path_group = parser.add_argument_group("Path and Logging Options")
     path_group.add_argument("--output-dir", default="./output", help="Output directory")
@@ -1100,6 +1113,8 @@ def main():
                 task=task,
                 features=features,
                 overrides=overrides,
+                device=args.device,
+                compute_type=args.compute_type,
             )
 
         elif args.mode == "transformers":
@@ -1114,6 +1129,8 @@ def main():
                 pipeline_name=args.mode,
                 sensitivity=args.sensitivity,
                 task=task,
+                device=args.device,
+                compute_type=args.compute_type,
             )
 
         if args.scene_detection_method:
@@ -1128,16 +1145,20 @@ def main():
     # Apply model override if specified via CLI (not for ensemble mode)
     if args.model and resolved_config is not None:
         logger.info(f"Overriding model with CLI argument: {args.model}")
+        # Determine device and compute_type for model override
+        # Priority: CLI args > auto-detection
+        override_device = args.device if args.device and args.device != "auto" else get_best_device()
+        override_compute_type = args.compute_type if args.compute_type and args.compute_type != "auto" else "int8"
         # Create a model configuration for the CLI-specified model
         override_model_config = {
             "provider": "openai_whisper",  # Default provider
             "model_name": args.model,
-            "device": get_best_device(),  # Auto-detect: CUDA -> MPS -> CPU
-            "compute_type": "int8",  # Default to int8 for quantized models (CTranslate2)
+            "device": override_device,
+            "compute_type": override_compute_type,
             "supported_tasks": ["transcribe", "translate"]
         }
         resolved_config["model"] = override_model_config
-        logger.debug(f"Model override applied: {args.model}")
+        logger.debug(f"Model override applied: {args.model} (device={override_device}, compute_type={override_compute_type})")
 
     # Apply language override to decoder params (not for ensemble mode)
     if resolved_config is not None and "params" in resolved_config and "decoder" in resolved_config["params"]:
@@ -1296,7 +1317,10 @@ def main():
                 'speech_enhancer': args.pass1_speech_enhancer,
                 'model': args.pass1_model,
                 'params': pass1_params,  # None = use defaults, object = custom
-                'hf_params': pass1_hf_params  # For transformers pipeline
+                'hf_params': pass1_hf_params,  # For transformers pipeline
+                'language': language_code,  # Source language code (e.g., 'en', 'ja')
+                'device': args.device,  # Hardware override (None = auto-detect)
+                'compute_type': args.compute_type,  # Compute type override (None = auto)
             }
 
             pass2_config = None
@@ -1309,7 +1333,10 @@ def main():
                     'speech_enhancer': args.pass2_speech_enhancer,
                     'model': args.pass2_model,
                     'params': pass2_params,
-                    'hf_params': pass2_hf_params  # For transformers pipeline
+                    'hf_params': pass2_hf_params,  # For transformers pipeline
+                    'language': language_code,  # Source language code (e.g., 'en', 'ja')
+                    'device': args.device,  # Hardware override (None = auto-detect)
+                    'compute_type': args.compute_type,  # Compute type override (None = auto)
                 }
 
             # Create orchestrator

whisperjav/modules/faster_whisper_pro_asr.py

@@ -47,21 +47,24 @@ class FasterWhisperProASR:
         # Use smart device detection: CUDA -> MPS -> CPU
         self.device = model_config.get("device", get_best_device())
 
-        # Smart compute_type default based on device (ctranslate2 backend)
-        # - CUDA: int8_float16 (quantized weights + FP16 tensor cores = fastest)
-        # - MPS:  float16 (safe for Apple Silicon)
-        # - CPU:  int8 (quantized weights + FP32 compute = only fast option)
-        requested_compute_type = model_config.get("compute_type")
-        if requested_compute_type is None or requested_compute_type == "auto":
-            if self.device == "cuda":
-                self.compute_type = "int8_float16"
-            elif self.device == "mps":
-                self.compute_type = "float16"
-            else:  # cpu
-                self.compute_type = "int8"
-            logger.debug(f"Auto-selected compute_type='{self.compute_type}' for device='{self.device}'")
-        else:
-            self.compute_type = requested_compute_type
+        # CTRANSLATE2 COMPATIBILITY: MPS is not supported by ctranslate2/faster-whisper
+        # CTranslate2 only supports "cuda" or "cpu" devices.
+        # See: https://github.com/OpenNMT/CTranslate2/issues/1562
+        # See: https://github.com/SYSTRAN/faster-whisper/issues/911
+        if self.device == "mps":
+            logger.warning(
+                "Apple Silicon MPS detected, but faster-whisper/ctranslate2 doesn't support MPS. "
+                "Using CPU mode with Apple Accelerate optimization. "
+                "For GPU acceleration on Mac, use --mode transformers instead."
+            )
+            self.device = "cpu"
+
+        # Use compute_type from resolver (defaults to "auto" for CTranslate2 providers)
+        # CTranslate2's "auto" handles device-specific optimization internally:
+        # - GPU capability detection (int8_float16 for RTX 20/30/40, float16 for RTX 50 Blackwell)
+        # - CPU instruction sets (AVX2, AVX512 → int8)
+        # See: https://github.com/OpenNMT/CTranslate2/issues/1865
+        self.compute_type = model_config.get("compute_type", "auto")
 
         decoder_params = params["decoder"]
         vad_params = params["vad"]
@@ -173,25 +176,7 @@ class FasterWhisperProASR:
     def _initialize_models(self):
         """Initialize Faster-Whisper model (direct API)."""
         # Note: Speech segmentation is handled by external Speech Segmenter (set in __init__)
-
-        # CPU COMPATIBILITY: float16/bfloat16 are NOT supported on CPU by ctranslate2
-        # Only float32 and int8 work on CPU. This check is ONLY for CPU - GPU paths unchanged.
-        # Note: Apple Silicon MPS (device="mps") DOES support float16/bfloat16 - this only affects CPU mode.
-        if self.device == "cpu" and self.compute_type in ("float16", "bfloat16", "int8_float16", "int8_bfloat16"):
-            import platform
-            original_compute_type = self.compute_type
-            self.compute_type = "int8"  # int8 is faster than float32 on CPU
-
-            # Tailor message based on platform
-            if platform.system() == "Darwin":
-                hint = "For GPU acceleration on Mac, ensure MPS is available (Apple Silicon required)."
-            else:
-                hint = "For GPU acceleration, use a CUDA-compatible GPU."
-
-            logger.warning(
-                f"Compute type '{original_compute_type}' is not supported on CPU. "
-                f"Automatically switching to 'int8' (best CPU performance). {hint}"
-            )
+        # Note: compute_type="auto" (default) lets CTranslate2 handle device-specific selection
 
         # Faster-Whisper direct API initialization
         logger.debug(

whisperjav/modules/kotoba_faster_whisper_asr.py

@@ -52,7 +52,22 @@ class KotobaFasterWhisperASR:
         # Model configuration
         self.model_name = model_config.get("model_name", self.DEFAULT_MODEL)
         self.device = model_config.get("device", get_best_device())
-        self.compute_type = model_config.get("compute_type", "float16")
+
+        # CTRANSLATE2 COMPATIBILITY: MPS is not supported by ctranslate2/faster-whisper
+        # CTranslate2 only supports "cuda" or "cpu" devices.
+        # See: https://github.com/OpenNMT/CTranslate2/issues/1562
+        if self.device == "mps":
+            logger.warning(
+                "Apple Silicon MPS detected, but faster-whisper/ctranslate2 doesn't support MPS. "
+                "Using CPU mode with Apple Accelerate optimization. "
+                "For GPU acceleration on Mac, use --mode transformers instead."
+            )
+            self.device = "cpu"
+
+        # Use compute_type from resolver (defaults to "auto" for CTranslate2 providers)
+        # CTranslate2's "auto" handles device-specific optimization internally
+        # See: https://github.com/OpenNMT/CTranslate2/issues/1865
+        self.compute_type = model_config.get("compute_type", "auto")
 
         # Task
         self.task = task
@@ -122,31 +137,22 @@ class KotobaFasterWhisperASR:
         logger.info(f"Loading Kotoba model: {self.model_name}")
         logger.info(f"Device: {self.device}, Compute type: {self.compute_type}")
 
-        # Determine compute type based on device capabilities
-        # On CPU, float16 is generally not supported by CTranslate2, so we fallback to int8
-        # unless float32 is explicitly requested.
-        if self.device == "cuda":
-            actual_compute_type = self.compute_type
-        else:
-            if self.compute_type == "float16":
-                logger.warning("float16 compute type is not supported on CPU. Falling back to int8.")
-                actual_compute_type = "int8"
-            else:
-                actual_compute_type = self.compute_type
+        # Note: compute_type="auto" (default) lets CTranslate2 handle device-specific selection
+        # See: https://github.com/OpenNMT/CTranslate2/issues/1865
 
         try:
             self.model = WhisperModel(
                 model_size_or_path=self.model_name,
                 device=self.device,
-                compute_type=actual_compute_type,
+                compute_type=self.compute_type,
                 download_root=None,  # Use default HuggingFace cache
                 cpu_threads=0,  # Use default
                 num_workers=1,
             )
             logger.info("Kotoba model loaded successfully")
         except Exception as e:
-            # C4: VRAM exhaustion fallback - try int8 if float16 fails on CUDA
-            if self.device == "cuda" and actual_compute_type == "float16":
+            # VRAM exhaustion fallback - try int8 if auto/float16 fails on CUDA
+            if self.device == "cuda" and self.compute_type != "int8":
                 vram_error_indicators = [
                     "out of memory",
                     "cuda out of memory",
@@ -158,7 +164,7 @@ class KotobaFasterWhisperASR:
                 error_str = str(e).lower()
                 if any(indicator in error_str for indicator in vram_error_indicators):
                     logger.warning(
-                        f"CUDA float16 failed (likely VRAM exhaustion): {e}. "
+                        f"CUDA {self.compute_type} failed (likely VRAM exhaustion): {e}. "
                         "Falling back to int8 compute type."
                     )
                     try:

whisperjav/modules/speech_segmentation/backends/whisper_vad.py

@@ -83,7 +83,7 @@ class WhisperVadSpeechSegmenter:
         no_speech_threshold: float = 0.6,
         logprob_threshold: float = -1.0,
         language: str = "ja",
-        compute_type: str = "float16",
+        compute_type: str = "auto",
         device: str = "auto",
         cache_results: bool = True,
         cache_dir: Optional[Path] = None,
@@ -265,16 +265,12 @@ class WhisperVadSpeechSegmenter:
             device = "cuda" if torch.cuda.is_available() else "cpu"
             logger.info(f"[WhisperVAD] Auto-detected device: {device}")
 
-        # Adjust compute_type for CPU
-        compute_type = self.compute_type
-        if device == "cpu" and compute_type in ("float16", "int8_float16"):
-            compute_type = "int8"
-            logger.info(f"[WhisperVAD] Adjusted compute_type for CPU: {compute_type}")
-
+        # Note: compute_type="auto" (default) lets CTranslate2 handle device-specific selection
+        # See: https://github.com/OpenNMT/CTranslate2/issues/1865
         self._model = WhisperModel(
             self._model_size,
             device=device,
-            compute_type=compute_type,
+            compute_type=self.compute_type,
         )
         logger.info(f"[WhisperVAD] Model loaded successfully on {device}")
 

whisperjav/modules/stable_ts_asr.py

@@ -148,9 +148,22 @@ class StableTSASR:
         self.model_name = model_config.get("model_name", "large-v2")
         # Use smart device detection: CUDA -> MPS -> CPU
         self.device = model_config.get("device", get_best_device())
+        self.turbo_mode = turbo_mode
+
+        # CTRANSLATE2 COMPATIBILITY: MPS is not supported by ctranslate2/faster-whisper
+        # In turbo_mode, stable-ts uses faster-whisper which uses ctranslate2.
+        # CTranslate2 only supports "cuda" or "cpu" devices.
+        # See: https://github.com/OpenNMT/CTranslate2/issues/1562
+        if self.device == "mps" and self.turbo_mode:
+            logger.warning(
+                "Apple Silicon MPS detected, but faster-whisper/ctranslate2 doesn't support MPS. "
+                "Using CPU mode with Apple Accelerate optimization. "
+                "For GPU acceleration on Mac, use --mode transformers instead."
+            )
+            self.device = "cpu"
+
         # Default to int8 for quantized models (faster-whisper uses CTranslate2 quantized models)
         self.compute_type = model_config.get("compute_type", "int8")
-        self.turbo_mode = turbo_mode
         self.model_repo = model_config.get("hf_repo")
         if not self.model_repo:
             self.model_repo = self._derive_model_repo(self.model_name)

whisperjav/version_checker.py

@@ -3,6 +3,7 @@
 WhisperJAV Version Checker
 
 Checks for new versions on GitHub and provides update notifications.
+Supports both stable releases and development (latest commit) updates.
 Uses GitHub API with caching and retry logic.
 """
 
@@ -10,11 +11,12 @@ import os
 import sys
 import json
 import time
+import subprocess
 import urllib.request
 import urllib.error
 from pathlib import Path
-from typing import Optional, Tuple, Dict
-from dataclasses import dataclass
+from typing import Optional, Tuple, Dict, List
+from dataclasses import dataclass, field
 from datetime import datetime, timedelta
 
 # Import version - handle both installed and dev scenarios
@@ -40,14 +42,23 @@ except ImportError:
 #   set WHISPERJAV_RELEASES_URL=http://localhost:8000/releases
 # =============================================================================
 
-# GitHub API endpoint
+# GitHub API endpoints
+GITHUB_REPO = os.environ.get('WHISPERJAV_REPO', 'meizhong986/whisperjav')
 GITHUB_API_URL = os.environ.get(
     'WHISPERJAV_UPDATE_API_URL',
-    'https://api.github.com/repos/meizhong986/whisperjav/releases/latest'
+    f'https://api.github.com/repos/{GITHUB_REPO}/releases/latest'
+)
+GITHUB_COMMITS_API_URL = os.environ.get(
+    'WHISPERJAV_COMMITS_API_URL',
+    f'https://api.github.com/repos/{GITHUB_REPO}/commits/main'
+)
+GITHUB_COMPARE_API_URL = os.environ.get(
+    'WHISPERJAV_COMPARE_API_URL',
+    f'https://api.github.com/repos/{GITHUB_REPO}/compare'
 )
 GITHUB_RELEASES_URL = os.environ.get(
     'WHISPERJAV_RELEASES_URL',
-    'https://github.com/meizhong986/WhisperJAV/releases'
+    f'https://github.com/{GITHUB_REPO}/releases'
 )
 
 # Cache settings
@@ -67,13 +78,44 @@ class VersionInfo:
     download_url: Optional[str] = None
 
 
+@dataclass
+class CommitInfo:
+    """Information about a single commit."""
+    sha: str
+    short_sha: str
+    message: str
+    author: str
+    date: str
+    url: str
+
+
+@dataclass
+class DevUpdateInfo:
+    """Information about development (latest commit) updates."""
+    commits_ahead: int
+    latest_commit: Optional[CommitInfo] = None
+    recent_commits: List[CommitInfo] = field(default_factory=list)
+    base_tag: Optional[str] = None  # The release tag we're comparing against
+
+
 @dataclass
 class UpdateCheckResult:
     """Result of an update check."""
-    update_available: bool
+    # Current installation info
     current_version: str
+    current_commit: Optional[str] = None  # Short SHA if known
+    installation_type: str = "release"  # "release" or "dev"
+
+    # Stable release track
+    update_available: bool = False
     latest_version: Optional[str] = None
     version_info: Optional[VersionInfo] = None
+
+    # Development track
+    dev_update_available: bool = False
+    dev_info: Optional[DevUpdateInfo] = None
+
+    # Meta
     error: Optional[str] = None
     from_cache: bool = False
 
@@ -207,17 +249,160 @@ def _fetch_latest_release(timeout: int = 10, retries: int = 2) -> Optional[Dict]
     return None
 
 
-def check_for_updates(force: bool = False) -> UpdateCheckResult:
+def _fetch_github_api(url: str, timeout: int = 10) -> Optional[Dict]:
+    """
+    Fetch data from GitHub API.
+
+    Args:
+        url: GitHub API URL
+        timeout: Request timeout in seconds
+
+    Returns:
+        JSON response dict or None on failure
+    """
+    try:
+        request = urllib.request.Request(
+            url,
+            headers={
+                'Accept': 'application/vnd.github.v3+json',
+                'User-Agent': f'WhisperJAV/{CURRENT_VERSION}'
+            }
+        )
+        with urllib.request.urlopen(request, timeout=timeout) as response:
+            return json.loads(response.read().decode('utf-8'))
+    except Exception:
+        return None
+
+
+def _get_current_commit() -> Optional[str]:
+    """
+    Get the current installation's commit hash.
+
+    Tries multiple methods:
+    1. Check if running from git repo (development)
+    2. Check stored commit hash from pip install
+
+    Returns:
+        Short commit hash (7 chars) or None
+    """
+    # Method 1: Running from git repo (development)
+    try:
+        result = subprocess.run(
+            ['git', 'rev-parse', '--short', 'HEAD'],
+            capture_output=True,
+            text=True,
+            timeout=5,
+            cwd=Path(__file__).parent.parent  # whisperjav package root
+        )
+        if result.returncode == 0:
+            return result.stdout.strip()[:7]
+    except Exception:
+        pass
+
+    # Method 2: Check for stored commit hash (from pip install git+...)
+    try:
+        commit_file = CACHE_DIR / 'installed_commit.txt'
+        if commit_file.exists():
+            return commit_file.read_text().strip()[:7]
+    except Exception:
+        pass
+
+    return None
+
+
+def _fetch_latest_commit(timeout: int = 10) -> Optional[CommitInfo]:
+    """
+    Fetch the latest commit from the main branch.
+
+    Returns:
+        CommitInfo or None on failure
+    """
+    data = _fetch_github_api(GITHUB_COMMITS_API_URL, timeout)
+    if not data:
+        return None
+
+    try:
+        commit = data.get('commit', {})
+        return CommitInfo(
+            sha=data.get('sha', ''),
+            short_sha=data.get('sha', '')[:7],
+            message=commit.get('message', '').split('\n')[0][:100],  # First line, truncated
+            author=commit.get('author', {}).get('name', 'Unknown'),
+            date=commit.get('author', {}).get('date', ''),
+            url=data.get('html_url', '')
+        )
+    except Exception:
+        return None
+
+
+def _fetch_commits_since_tag(tag: str, timeout: int = 10) -> Optional[DevUpdateInfo]:
+    """
+    Fetch commits between a release tag and the main branch.
+
+    Args:
+        tag: Release tag to compare from (e.g., 'v1.7.5')
+
+    Returns:
+        DevUpdateInfo with commit list, or None on failure
+    """
+    # Use GitHub compare API: /repos/{owner}/{repo}/compare/{base}...{head}
+    compare_url = f"{GITHUB_COMPARE_API_URL}/{tag}...main"
+    data = _fetch_github_api(compare_url, timeout)
+
+    if not data:
+        return None
+
+    try:
+        commits_data = data.get('commits', [])
+        ahead_by = data.get('ahead_by', len(commits_data))
+
+        # Parse recent commits (up to 10)
+        recent_commits = []
+        for commit_data in commits_data[-10:]:  # Last 10 commits
+            commit = commit_data.get('commit', {})
+            recent_commits.append(CommitInfo(
+                sha=commit_data.get('sha', ''),
+                short_sha=commit_data.get('sha', '')[:7],
+                message=commit.get('message', '').split('\n')[0][:80],
+                author=commit.get('author', {}).get('name', 'Unknown'),
+                date=commit.get('author', {}).get('date', ''),
+                url=commit_data.get('html_url', '')
+            ))
+
+        # Reverse to show newest first
+        recent_commits.reverse()
+
+        return DevUpdateInfo(
+            commits_ahead=ahead_by,
+            latest_commit=recent_commits[0] if recent_commits else None,
+            recent_commits=recent_commits,
+            base_tag=tag
+        )
+    except Exception:
+        return None
+
+
+def check_for_updates(force: bool = False, include_dev: bool = True) -> UpdateCheckResult:
     """
     Check if a newer version of WhisperJAV is available.
 
+    Checks both:
+    1. Stable releases (tagged versions on GitHub)
+    2. Development updates (commits on main branch since last release)
+
     Args:
         force: If True, bypass cache and check immediately
+        include_dev: If True, also check for dev updates (commits since release)
 
     Returns:
-        UpdateCheckResult with update information
+        UpdateCheckResult with both release and dev update information
     """
     current = CURRENT_VERSION
+    current_commit = _get_current_commit()
+
+    # Determine installation type
+    # If we can get a commit hash, likely installed from git
+    installation_type = "dev" if current_commit else "release"
 
     # Check cache first (unless forced)
     if not force:
@@ -239,31 +424,76 @@ def check_for_updates(force: bool = False) -> UpdateCheckResult:
                         download_url=vi.get('download_url')
                     )
 
+                # Reconstruct dev_info from cache if available
+                dev_info = None
+                dev_update_available = False
+                if cached.get('dev_info'):
+                    di = cached['dev_info']
+                    commits_ahead = di.get('commits_ahead', 0)
+                    dev_update_available = commits_ahead > 0
+
+                    recent_commits = []
+                    for c in di.get('recent_commits', []):
+                        recent_commits.append(CommitInfo(
+                            sha=c.get('sha', ''),
+                            short_sha=c.get('short_sha', ''),
+                            message=c.get('message', ''),
+                            author=c.get('author', ''),
+                            date=c.get('date', ''),
+                            url=c.get('url', '')
+                        ))
+
+                    latest_commit = None
+                    if di.get('latest_commit'):
+                        lc = di['latest_commit']
+                        latest_commit = CommitInfo(
+                            sha=lc.get('sha', ''),
+                            short_sha=lc.get('short_sha', ''),
+                            message=lc.get('message', ''),
+                            author=lc.get('author', ''),
+                            date=lc.get('date', ''),
+                            url=lc.get('url', '')
+                        )
+
+                    dev_info = DevUpdateInfo(
+                        commits_ahead=commits_ahead,
+                        latest_commit=latest_commit,
+                        recent_commits=recent_commits,
+                        base_tag=di.get('base_tag')
+                    )
+
                 return UpdateCheckResult(
-                    update_available=update_available,
                     current_version=current,
+                    current_commit=current_commit,
+                    installation_type=installation_type,
+                    update_available=update_available,
                     latest_version=latest,
                     version_info=version_info,
+                    dev_update_available=dev_update_available,
+                    dev_info=dev_info,
                     from_cache=True
                 )
 
-    # Fetch from GitHub
+    # Fetch latest release from GitHub
     release_data = _fetch_latest_release()
 
     if not release_data:
         return UpdateCheckResult(
-            update_available=False,
             current_version=current,
+            current_commit=current_commit,
+            installation_type=installation_type,
             error="Could not check for updates"
         )
 
     # Parse release data
     latest = release_data.get('tag_name', '').lstrip('v')
+    release_tag = release_data.get('tag_name', f'v{latest}')
 
     if not latest:
         return UpdateCheckResult(
-            update_available=False,
             current_version=current,
+            current_commit=current_commit,
+            installation_type=installation_type,
             error="Invalid release data"
         )
 
@@ -283,6 +513,26 @@ def check_for_updates(force: bool = False) -> UpdateCheckResult:
             version_info.download_url = asset.get('browser_download_url')
             break
 
+    # Compare versions for stable track
+    update_available = compare_versions(current, latest) < 0
+
+    # Fetch dev track info (commits since release)
+    dev_info = None
+    dev_update_available = False
+
+    if include_dev:
+        dev_info = _fetch_commits_since_tag(release_tag)
+        if dev_info:
+            # Dev update available if there are commits ahead
+            # AND user doesn't already have the latest commit
+            if dev_info.commits_ahead > 0:
+                if current_commit and dev_info.latest_commit:
+                    # User has a commit hash - check if it matches latest
+                    dev_update_available = current_commit != dev_info.latest_commit.short_sha
+                else:
+                    # Can't compare commits, assume update available
+                    dev_update_available = True
+
     # Save to cache
     cache_data = {
         'latest_version': latest,
@@ -295,16 +545,44 @@ def check_for_updates(force: bool = False) -> UpdateCheckResult:
             'download_url': version_info.download_url
         }
     }
+
+    # Cache dev info if available
+    if dev_info:
+        cache_data['dev_info'] = {
+            'commits_ahead': dev_info.commits_ahead,
+            'base_tag': dev_info.base_tag,
+            'latest_commit': {
+                'sha': dev_info.latest_commit.sha,
+                'short_sha': dev_info.latest_commit.short_sha,
+                'message': dev_info.latest_commit.message,
+                'author': dev_info.latest_commit.author,
+                'date': dev_info.latest_commit.date,
+                'url': dev_info.latest_commit.url
+            } if dev_info.latest_commit else None,
+            'recent_commits': [
+                {
+                    'sha': c.sha,
+                    'short_sha': c.short_sha,
+                    'message': c.message,
+                    'author': c.author,
+                    'date': c.date,
+                    'url': c.url
+                }
+                for c in dev_info.recent_commits[:5]  # Cache up to 5 commits
+            ]
+        }
+
     _save_cache(cache_data)
 
-    # Compare versions
-    update_available = compare_versions(current, latest) < 0
-
     return UpdateCheckResult(
-        update_available=update_available,
         current_version=current,
+        current_commit=current_commit,
+        installation_type=installation_type,
+        update_available=update_available,
         latest_version=latest,
         version_info=version_info,
+        dev_update_available=dev_update_available,
+        dev_info=dev_info,
         from_cache=False
     )
 
@@ -385,16 +663,52 @@ def should_show_notification(level: str, last_dismissed: Optional[str] = None) -
 # CLI interface
 if __name__ == "__main__":
     print(f"Current version: {CURRENT_VERSION}")
-    print("Checking for updates...")
+    if _get_current_commit():
+        print(f"Current commit: {_get_current_commit()}")
+    print("Checking for updates...\n")
 
     result = check_for_updates(force=True)
 
     if result.error:
         print(f"Error: {result.error}")
-    elif result.update_available:
-        print(f"Update available: {result.latest_version}")
-        print(f"Release URL: {result.version_info.release_url}")
+        sys.exit(1)
+
+    # Stable release track
+    print("=" * 50)
+    print("[STABLE RELEASE]")
+    print("=" * 50)
+    if result.update_available:
+        print(f"  Update available: v{result.latest_version}")
+        print(f"  Release URL: {result.version_info.release_url}")
         if result.version_info.download_url:
-            print(f"Download: {result.version_info.download_url}")
+            print(f"  Download: {result.version_info.download_url}")
     else:
-        print("You have the latest version!")
+        print(f"  [OK] You have the latest stable release (v{result.latest_version})")
+
+    # Development track
+    print("\n" + "=" * 50)
+    print("[DEVELOPMENT] (main branch)")
+    print("=" * 50)
+    if result.dev_info:
+        if result.dev_info.commits_ahead > 0:
+            print(f"  {result.dev_info.commits_ahead} commit(s) ahead of {result.dev_info.base_tag}")
+            if result.dev_update_available:
+                print("  -> Updates available!")
+            else:
+                print("  [OK] You have the latest development version")
+
+            if result.dev_info.recent_commits:
+                print("\n  Recent commits:")
+                for commit in result.dev_info.recent_commits[:5]:
+                    # Encode safely for Windows console
+                    msg = commit.message[:60].encode('ascii', 'replace').decode('ascii')
+                    print(f"    * [{commit.short_sha}] {msg}")
+        else:
+            print("  [OK] Release is up to date with main branch")
+    else:
+        print("  Could not fetch development info")
+
+    print("\n" + "=" * 50)
+    print("To update:")
+    print("  Stable:  pip install -U whisperjav")
+    print("  Dev:     pip install -U git+https://github.com/meizhong986/whisperjav.git")

whisperjav/webview_gui/api.py

@@ -12,6 +12,7 @@ import time
 import queue
 import threading
 import subprocess
+import webbrowser
 from pathlib import Path
 from typing import Optional, List, Dict, Any
 import webview
@@ -2104,17 +2105,30 @@ class WhisperJAVAPI:
         """
         Check if a newer version of WhisperJAV is available.
 
+        Checks both stable releases and development (latest commit) updates.
+
         Args:
             force: If True, bypass cache and check immediately
 
         Returns:
             dict: Update check result with:
-                - update_available: bool
+                - success: bool
                 - current_version: str
+                - current_commit: str or None (short SHA)
+                - installation_type: "release" or "dev"
+
+                Stable track:
+                - update_available: bool
                 - latest_version: str (if available)
                 - notification_level: str (critical/major/minor/patch/none)
                 - release_url: str (if update available)
                 - release_notes: str (if update available)
+
+                Development track:
+                - dev_update_available: bool
+                - dev_commits_ahead: int
+                - dev_recent_commits: list of {sha, message, author, date}
+
                 - error: str (if check failed)
         """
         try:
@@ -2123,25 +2137,55 @@ class WhisperJAVAPI:
                 get_update_notification_level,
             )
 
-            result = check_for_updates(force=force)
+            result = check_for_updates(force=force, include_dev=True)
 
             response = {
                 "success": True,
-                "update_available": result.update_available,
+                # Current installation info
                 "current_version": result.current_version,
+                "current_commit": result.current_commit,
+                "installation_type": result.installation_type,
+                # Stable release track
+                "update_available": result.update_available,
                 "latest_version": result.latest_version,
                 "from_cache": result.from_cache,
+                # Development track
+                "dev_update_available": result.dev_update_available,
             }
 
             if result.error:
                 response["error"] = result.error
 
+            # Stable release info
             if result.version_info:
                 response["notification_level"] = get_update_notification_level(result)
                 response["release_url"] = result.version_info.release_url
                 response["release_notes"] = result.version_info.release_notes[:500]  # Truncate
                 response["is_prerelease"] = result.version_info.is_prerelease
 
+            # Development track info
+            if result.dev_info:
+                response["dev_commits_ahead"] = result.dev_info.commits_ahead
+                response["dev_base_tag"] = result.dev_info.base_tag
+                response["dev_recent_commits"] = [
+                    {
+                        "sha": c.short_sha,
+                        "message": c.message,
+                        "author": c.author,
+                        "date": c.date,
+                        "url": c.url
+                    }
+                    for c in result.dev_info.recent_commits[:5]
+                ]
+                if result.dev_info.latest_commit:
+                    response["dev_latest_commit"] = {
+                        "sha": result.dev_info.latest_commit.short_sha,
+                        "message": result.dev_info.latest_commit.message,
+                        "author": result.dev_info.latest_commit.author,
+                        "date": result.dev_info.latest_commit.date,
+                        "url": result.dev_info.latest_commit.url
+                    }
+
             return response
 
         except Exception as e:
@@ -2149,6 +2193,7 @@ class WhisperJAVAPI:
                 "success": False,
                 "error": str(e),
                 "update_available": False,
+                "dev_update_available": False,
                 "current_version": "unknown"
             }
 
@@ -2274,3 +2319,19 @@ class WhisperJAVAPI:
                 "error": str(e),
                 "should_exit": False
             }
+
+    def open_url(self, url: str) -> Dict[str, Any]:
+        """
+        Open a URL in the system's default web browser.
+
+        Args:
+            url: The URL to open
+
+        Returns:
+            dict: Success status
+        """
+        try:
+            webbrowser.open(url)
+            return {"success": True}
+        except Exception as e:
+            return {"success": False, "error": str(e)}

whisperjav/webview_gui/assets/app.js

@@ -4076,6 +4076,240 @@ const UpdateManager = {
     }
 };
 
+// ============================================================
+// Update Check Manager (manual update check from menu)
+// ============================================================
+const UpdateCheckManager = {
+    modal: null,
+
+    init() {
+        this.modal = document.getElementById('updateCheckModal');
+        if (!this.modal) {
+            console.warn('UpdateCheckManager: modal not found');
+            return;
+        }
+
+        // Menu item click
+        const checkBtn = document.getElementById('checkUpdatesBtn');
+        if (checkBtn) {
+            checkBtn.addEventListener('click', (e) => {
+                e.stopPropagation();
+                document.getElementById('themeMenu')?.classList.remove('active');
+                this.show();
+            });
+        }
+
+        // Modal close handlers
+        document.getElementById('updateCheckModalClose')?.addEventListener('click', () => this.close());
+        document.getElementById('updateCheckLater')?.addEventListener('click', () => this.close());
+        document.getElementById('updateCheckNow')?.addEventListener('click', () => this.startUpdate());
+        document.getElementById('updateCheckDownload')?.addEventListener('click', () => this.openDownloadPage());
+
+        // Close on overlay click
+        this.modal.addEventListener('click', (e) => {
+            if (e.target === this.modal) this.close();
+        });
+
+        // Close on Escape key
+        document.addEventListener('keydown', (e) => {
+            if (e.key === 'Escape' && this.modal.classList.contains('active')) {
+                this.close();
+            }
+        });
+    },
+
+    async show() {
+        // Reset and show loading state
+        this.showLoading();
+        this.modal.classList.add('active');
+
+        try {
+            // Force fresh check (bypass cache)
+            const result = await pywebview.api.check_for_updates(true);
+            this.showResult(result);
+        } catch (err) {
+            console.error('Update check error:', err);
+            this.showError(err);
+        }
+    },
+
+    // Store the last result for update actions
+    lastResult: null,
+
+    showLoading() {
+        document.getElementById('updateCheckLoading').style.display = 'block';
+        document.getElementById('updateCheckResult').style.display = 'none';
+        document.getElementById('updateCheckError').style.display = 'none';
+    },
+
+    showResult(result) {
+        this.lastResult = result;
+        document.getElementById('updateCheckLoading').style.display = 'none';
+        document.getElementById('updateCheckResult').style.display = 'block';
+
+        // Set current version info
+        document.getElementById('updateCurrentVersion').textContent = 'v' + result.current_version;
+
+        // Show commit hash if available
+        const commitEl = document.getElementById('updateCurrentCommit');
+        if (commitEl && result.current_commit) {
+            commitEl.textContent = '(' + result.current_commit.substring(0, 7) + ')';
+            commitEl.style.display = 'inline';
+        } else if (commitEl) {
+            commitEl.style.display = 'none';
+        }
+
+        // === Stable Release Track ===
+        this.updateStableTrack(result);
+
+        // === Development Track ===
+        this.updateDevTrack(result);
+    },
+
+    updateStableTrack(result) {
+        const stableUpToDate = document.getElementById('stableUpToDate');
+        const stableUpdateAvailable = document.getElementById('stableUpdateAvailable');
+        const stableReleaseNotes = document.getElementById('stableReleaseNotes');
+        const stableActions = document.getElementById('stableActions');
+        const stableLatestVersion = document.getElementById('stableLatestVersion');
+        const stableNewVersion = document.getElementById('stableNewVersion');
+        const stableBadge = document.getElementById('stableUpdateBadge');
+
+        if (result.update_available && result.latest_version) {
+            // Stable update available
+            stableUpToDate.style.display = 'none';
+            stableUpdateAvailable.style.display = 'flex';
+            stableNewVersion.textContent = result.latest_version;
+
+            // Set badge color based on notification level
+            const level = result.notification_level || 'patch';
+            stableBadge.textContent = level.toUpperCase();
+            stableBadge.className = 'update-badge ' + level;
+
+            // Show release notes if available
+            if (result.release_notes) {
+                const notesContent = document.getElementById('stableReleaseNotesContent');
+                notesContent.innerHTML = this.parseMarkdown(result.release_notes);
+                stableReleaseNotes.style.display = 'block';
+            } else {
+                stableReleaseNotes.style.display = 'none';
+            }
+
+            // Show action button (Download for major, Update for others)
+            if (level === 'major') {
+                stableActions.innerHTML = '<button id="updateToStable" class="btn btn-primary btn-sm">Download from GitHub</button>';
+                document.getElementById('updateToStable').addEventListener('click', () => this.openDownloadPage());
+            } else {
+                stableActions.innerHTML = '<button id="updateToStable" class="btn btn-primary btn-sm">Update to Stable</button>';
+                document.getElementById('updateToStable').addEventListener('click', () => this.startUpdate('stable'));
+            }
+            stableActions.style.display = 'block';
+        } else {
+            // Up to date with stable
+            stableUpToDate.style.display = 'flex';
+            stableLatestVersion.textContent = result.current_version;
+            stableUpdateAvailable.style.display = 'none';
+            stableReleaseNotes.style.display = 'none';
+            stableActions.style.display = 'none';
+        }
+    },
+
+    updateDevTrack(result) {
+        const devUpToDate = document.getElementById('devUpToDate');
+        const devUpdateAvailable = document.getElementById('devUpdateAvailable');
+        const devCommitsList = document.getElementById('devCommitsList');
+        const devActions = document.getElementById('devActions');
+        const devCommitsAhead = document.getElementById('devCommitsAhead');
+        const devRecentCommits = document.getElementById('devRecentCommits');
+
+        if (result.dev_update_available && result.dev_commits_ahead > 0) {
+            // Dev updates available
+            devUpToDate.style.display = 'none';
+            devUpdateAvailable.style.display = 'flex';
+            devCommitsAhead.textContent = result.dev_commits_ahead;
+
+            // Show recent commits list
+            if (result.dev_recent_commits && result.dev_recent_commits.length > 0) {
+                devRecentCommits.innerHTML = result.dev_recent_commits
+                    .slice(0, 5) // Show max 5 recent commits
+                    .map(commit => {
+                        const shortSha = commit.short_sha || commit.sha.substring(0, 7);
+                        const message = this.escapeHtml(commit.message.split('\n')[0]); // First line only
+                        return `<li><code>${shortSha}</code> ${message}</li>`;
+                    })
+                    .join('');
+                devCommitsList.style.display = 'block';
+            } else {
+                devCommitsList.style.display = 'none';
+            }
+
+            // Show action button
+            devActions.innerHTML = '<button id="updateToDev" class="btn btn-secondary btn-sm">Update to Latest Dev</button>';
+            document.getElementById('updateToDev').addEventListener('click', () => this.startUpdate('dev'));
+            devActions.style.display = 'block';
+        } else {
+            // Up to date with dev
+            devUpToDate.style.display = 'flex';
+            devUpdateAvailable.style.display = 'none';
+            devCommitsList.style.display = 'none';
+            devActions.style.display = 'none';
+        }
+    },
+
+    showError(error) {
+        document.getElementById('updateCheckLoading').style.display = 'none';
+        document.getElementById('updateCheckResult').style.display = 'none';
+        document.getElementById('updateCheckError').style.display = 'block';
+    },
+
+    async startUpdate(track = 'stable') {
+        // Close modal
+        this.close();
+
+        // Use existing UpdateManager to show progress and start update
+        // Both tracks use the same update mechanism (pip install from GitHub)
+        // The difference is just which version was chosen
+        console.log('Starting update from track:', track);
+        UpdateManager.startUpdate();
+    },
+
+    async openDownloadPage() {
+        // Open GitHub releases in browser
+        try {
+            await pywebview.api.open_url('https://github.com/meizhong986/WhisperJAV/releases');
+        } catch (err) {
+            // Fallback: open in new window
+            window.open('https://github.com/meizhong986/WhisperJAV/releases', '_blank');
+        }
+        this.close();
+    },
+
+    close() {
+        this.modal.classList.remove('active');
+    },
+
+    escapeHtml(text) {
+        const div = document.createElement('div');
+        div.textContent = text;
+        return div.innerHTML;
+    },
+
+    parseMarkdown(text) {
+        // Simple markdown to HTML conversion
+        if (!text) return '';
+        return text
+            .replace(/^### (.+)$/gm, '<h4>$1</h4>')
+            .replace(/^## (.+)$/gm, '<h3>$1</h3>')
+            .replace(/^# (.+)$/gm, '<h2>$1</h2>')
+            .replace(/\*\*(.+?)\*\*/g, '<strong>$1</strong>')
+            .replace(/`([^`]+)`/g, '<code>$1</code>')
+            .replace(/^\* (.+)$/gm, '<li>$1</li>')
+            .replace(/^- (.+)$/gm, '<li>$1</li>')
+            .replace(/(<li>.*<\/li>)/s, '<ul>$1</ul>')
+            .replace(/\n\n/g, '<br><br>');
+    }
+};
+
 // ============================================================
 // Initialization
 // ============================================================
@@ -4098,6 +4332,7 @@ document.addEventListener('DOMContentLoaded', async () => {
     ThemeManager.init();
     EnsembleManager.init();
     UpdateManager.init();
+    UpdateCheckManager.init();
 
     // Initial validation
     FormManager.validateForm();

whisperjav/webview_gui/assets/index.html

@@ -25,6 +25,10 @@
                     <button class="theme-option" data-theme="google" role="menuitem">Google Theme</button>
                     <button class="theme-option" data-theme="carbon" role="menuitem">IBM Carbon Theme</button>
                     <button class="theme-option" data-theme="primer" role="menuitem">GitHub Primer Theme</button>
+                    <div class="theme-menu-divider"></div>
+                    <button class="theme-option update-check-option" id="checkUpdatesBtn" role="menuitem">
+                        <span class="update-check-icon">&#x2B06;</span> Check for Updates
+                    </button>
                 </div>
             </div>
         </header>
@@ -838,6 +842,95 @@
         </div>
     </div>
 
+    <!-- Update Check Modal -->
+    <div class="modal-overlay" id="updateCheckModal">
+        <div class="modal-dialog modal-update-check">
+            <div class="modal-header">
+                <h2>Check for Updates</h2>
+                <button class="modal-close" id="updateCheckModalClose" aria-label="Close">&times;</button>
+            </div>
+            <div class="modal-body">
+                <!-- Loading State -->
+                <div id="updateCheckLoading" class="update-check-loading">
+                    <div class="spinner"></div>
+                    <p>Checking for updates...</p>
+                </div>
+
+                <!-- Result State -->
+                <div id="updateCheckResult" class="update-check-result" style="display: none;">
+                    <!-- Current Version Info -->
+                    <div class="current-version-info">
+                        <span class="version-label">Installed:</span>
+                        <span class="version-value" id="updateCurrentVersion">v1.7.5</span>
+                        <span class="commit-hash" id="updateCurrentCommit"></span>
+                    </div>
+
+                    <!-- Stable Release Track -->
+                    <div class="update-track update-track-stable" id="stableTrack">
+                        <div class="track-header">
+                            <span class="track-icon">&#x1F4E6;</span>
+                            <span class="track-title">Stable Release</span>
+                        </div>
+                        <div class="track-content">
+                            <div id="stableUpToDate" class="track-status track-status-ok" style="display: none;">
+                                <span class="status-icon">&#x2713;</span>
+                                <span>v<span id="stableLatestVersion">1.7.5</span> - You're up to date</span>
+                            </div>
+                            <div id="stableUpdateAvailable" class="track-status track-status-update" style="display: none;">
+                                <span class="status-icon">&#x2191;</span>
+                                <span>v<span id="stableNewVersion">1.7.6</span> available</span>
+                                <span class="update-badge" id="stableUpdateBadge">PATCH</span>
+                            </div>
+                            <div id="stableReleaseNotes" class="release-notes-section" style="display: none;">
+                                <details class="release-notes-details">
+                                    <summary>Release Notes</summary>
+                                    <div id="stableReleaseNotesContent" class="release-notes-content"></div>
+                                </details>
+                            </div>
+                            <div id="stableActions" class="track-actions" style="display: none;">
+                                <button id="updateToStable" class="btn btn-primary btn-sm">Update to Stable</button>
+                            </div>
+                        </div>
+                    </div>
+
+                    <!-- Development Track -->
+                    <div class="update-track update-track-dev" id="devTrack">
+                        <div class="track-header">
+                            <span class="track-icon">&#x26A1;</span>
+                            <span class="track-title">Development</span>
+                            <span class="track-subtitle">(main branch)</span>
+                        </div>
+                        <div class="track-content">
+                            <div id="devUpToDate" class="track-status track-status-ok" style="display: none;">
+                                <span class="status-icon">&#x2713;</span>
+                                <span>Up to date with main branch</span>
+                            </div>
+                            <div id="devUpdateAvailable" class="track-status track-status-update" style="display: none;">
+                                <span class="status-icon">&#x2191;</span>
+                                <span><span id="devCommitsAhead">0</span> commit(s) ahead</span>
+                            </div>
+                            <div id="devCommitsList" class="dev-commits-list" style="display: none;">
+                                <div class="commits-header">Recent commits:</div>
+                                <ul id="devRecentCommits" class="commits-list"></ul>
+                            </div>
+                            <div id="devActions" class="track-actions" style="display: none;">
+                                <button id="updateToDev" class="btn btn-secondary btn-sm">Update to Latest Dev</button>
+                            </div>
+                        </div>
+                    </div>
+                </div>
+
+                <!-- Error State -->
+                <div id="updateCheckError" class="update-check-error" style="display: none;">
+                    <p>Failed to check for updates. Please check your internet connection and try again.</p>
+                </div>
+            </div>
+            <div class="modal-footer" id="updateCheckFooter">
+                <button id="updateCheckLater" class="btn btn-secondary">Close</button>
+            </div>
+        </div>
+    </div>
+
     <!-- Customize Component Modal -->
     <div class="modal-overlay" id="customizeModal">
         <div class="modal-dialog modal-lg">

whisperjav/webview_gui/assets/style.css

@@ -2405,3 +2405,379 @@ button:focus-visible,
 .changelog-error p {
     margin-bottom: 16px;
 }
+
+/* ============================================================
+   Update Check Modal
+   ============================================================ */
+
+/* Theme menu divider */
+.theme-menu-divider {
+    height: 1px;
+    background: var(--border-color);
+    margin: 4px 0;
+}
+
+/* Check for Updates menu item */
+.update-check-option {
+    font-weight: 600;
+}
+
+.update-check-icon {
+    margin-right: 6px;
+}
+
+/* Update check loading state */
+.update-check-loading {
+    text-align: center;
+    padding: 40px 20px;
+}
+
+.update-check-loading .spinner {
+    width: 40px;
+    height: 40px;
+    border: 3px solid var(--border-color);
+    border-top-color: var(--primary-color);
+    border-radius: 50%;
+    animation: spin 1s linear infinite;
+    margin: 0 auto 16px;
+}
+
+.update-check-loading p {
+    color: var(--text-secondary);
+    margin: 0;
+}
+
+/* Result container */
+.update-check-result {
+    padding: 20px;
+}
+
+/* Version info display */
+.version-info {
+    margin-bottom: 16px;
+}
+
+.version-row {
+    display: flex;
+    align-items: center;
+    gap: 8px;
+    font-size: 16px;
+    margin: 8px 0;
+}
+
+.version-label {
+    color: var(--text-secondary);
+    min-width: 70px;
+}
+
+.version-value {
+    font-weight: 500;
+}
+
+.version-current .version-value {
+    color: var(--text-secondary);
+}
+
+.version-latest .version-value {
+    color: var(--text-primary);
+}
+
+/* Update type badges */
+.update-badge {
+    display: inline-block;
+    padding: 2px 8px;
+    border-radius: 4px;
+    font-size: 11px;
+    font-weight: 600;
+    margin-left: 8px;
+    text-transform: uppercase;
+}
+
+.update-badge.patch {
+    background: #28a745;
+    color: white;
+}
+
+.update-badge.minor {
+    background: #17a2b8;
+    color: white;
+}
+
+.update-badge.major {
+    background: #fd7e14;
+    color: white;
+}
+
+.update-badge.critical {
+    background: #dc3545;
+    color: white;
+}
+
+/* Up to date message */
+.update-up-to-date {
+    display: flex;
+    flex-direction: column;
+    align-items: center;
+    justify-content: center;
+    padding: 30px 20px;
+    color: #28a745;
+    font-size: 18px;
+}
+
+.update-up-to-date .checkmark {
+    font-size: 48px;
+    margin-bottom: 8px;
+}
+
+/* Release notes */
+.release-notes-details {
+    margin-top: 16px;
+    border: 1px solid var(--border-color);
+    border-radius: var(--border-radius);
+}
+
+.release-notes-details summary {
+    padding: 12px;
+    cursor: pointer;
+    background: var(--hover-tab-bg);
+    font-weight: 500;
+}
+
+.release-notes-details summary:hover {
+    background: var(--disabled-bg);
+}
+
+.release-notes-content {
+    padding: 12px;
+    max-height: 200px;
+    overflow-y: auto;
+    font-size: 13px;
+    line-height: 1.5;
+}
+
+.release-notes-content h3,
+.release-notes-content h4 {
+    margin: 12px 0 8px;
+    font-size: 14px;
+}
+
+.release-notes-content h3:first-child,
+.release-notes-content h4:first-child {
+    margin-top: 0;
+}
+
+.release-notes-content li {
+    margin: 4px 0;
+    padding-left: 4px;
+}
+
+.release-notes-content code {
+    background: var(--disabled-bg);
+    padding: 1px 4px;
+    border-radius: 3px;
+    font-family: var(--font-mono);
+    font-size: 12px;
+}
+
+/* Error state */
+.update-check-error {
+    text-align: center;
+    padding: 30px 20px;
+    color: var(--text-secondary);
+}
+
+.update-check-error p {
+    margin: 0;
+}
+
+/* ============================================================
+   Dual-Track Update Modal Styles
+   ============================================================ */
+
+/* Modal sizing */
+.modal-update-check {
+    max-width: 520px;
+    width: 90%;
+}
+
+/* Current version info at top */
+.current-version-info {
+    display: flex;
+    align-items: center;
+    gap: 8px;
+    padding: 12px 16px;
+    background: var(--inactive-tab-bg);
+    border-radius: var(--border-radius);
+    margin-bottom: 16px;
+    font-size: 14px;
+}
+
+.current-version-info .version-label {
+    color: var(--text-secondary);
+}
+
+.current-version-info .version-value {
+    font-weight: 600;
+    color: var(--text-primary);
+}
+
+.current-version-info .commit-hash {
+    font-family: var(--font-mono);
+    font-size: 12px;
+    color: var(--text-muted);
+}
+
+/* Update track containers */
+.update-track {
+    border: 1px solid var(--border-color);
+    border-radius: var(--border-radius-lg);
+    margin-bottom: 12px;
+    overflow: hidden;
+}
+
+.update-track:last-child {
+    margin-bottom: 0;
+}
+
+/* Track header */
+.track-header {
+    display: flex;
+    align-items: center;
+    gap: 8px;
+    padding: 10px 14px;
+    background: var(--inactive-tab-bg);
+    border-bottom: 1px solid var(--border-color);
+}
+
+.track-icon {
+    font-size: 16px;
+    line-height: 1;
+}
+
+.track-title {
+    font-weight: 600;
+    font-size: 13px;
+    color: var(--text-primary);
+}
+
+.track-subtitle {
+    font-size: 11px;
+    color: var(--text-muted);
+    font-weight: normal;
+}
+
+/* Stable track specific colors */
+.update-track-stable .track-header {
+    background: linear-gradient(135deg, rgba(40, 167, 69, 0.08) 0%, rgba(40, 167, 69, 0.04) 100%);
+}
+
+.update-track-stable .track-icon {
+    color: #28a745;
+}
+
+/* Dev track specific colors */
+.update-track-dev .track-header {
+    background: linear-gradient(135deg, rgba(253, 126, 20, 0.08) 0%, rgba(253, 126, 20, 0.04) 100%);
+}
+
+.update-track-dev .track-icon {
+    color: #fd7e14;
+}
+
+/* Track content */
+.track-content {
+    padding: 12px 14px;
+}
+
+/* Track status line */
+.track-status {
+    display: flex;
+    align-items: center;
+    gap: 8px;
+    font-size: 13px;
+}
+
+.track-status .status-icon {
+    font-size: 14px;
+    font-weight: bold;
+}
+
+.track-status-ok {
+    color: #28a745;
+}
+
+.track-status-ok .status-icon {
+    color: #28a745;
+}
+
+.track-status-update {
+    color: var(--text-primary);
+}
+
+.track-status-update .status-icon {
+    color: #fd7e14;
+}
+
+/* Release notes section */
+.release-notes-section {
+    margin-top: 10px;
+}
+
+/* Track actions */
+.track-actions {
+    margin-top: 12px;
+    padding-top: 10px;
+    border-top: 1px solid var(--border-color);
+}
+
+/* Dev commits list */
+.dev-commits-list {
+    margin-top: 10px;
+    padding: 10px;
+    background: var(--inactive-tab-bg);
+    border-radius: var(--border-radius);
+}
+
+.commits-header {
+    font-size: 11px;
+    font-weight: 600;
+    color: var(--text-secondary);
+    text-transform: uppercase;
+    letter-spacing: 0.3px;
+    margin-bottom: 8px;
+}
+
+.commits-list {
+    list-style: none;
+    padding: 0;
+    margin: 0;
+}
+
+.commits-list li {
+    font-size: 12px;
+    padding: 4px 0;
+    border-bottom: 1px solid var(--border-color);
+    line-height: 1.4;
+    display: flex;
+    gap: 8px;
+    align-items: flex-start;
+}
+
+.commits-list li:last-child {
+    border-bottom: none;
+}
+
+.commits-list code {
+    background: var(--disabled-bg);
+    padding: 2px 5px;
+    border-radius: 3px;
+    font-family: var(--font-mono);
+    font-size: 10px;
+    color: var(--primary-color);
+    flex-shrink: 0;
+}
+
+/* Ensure spinner shows properly in loading state */
+.update-check-loading .spinner {
+    display: block;
+}