ch-srtgen 1.0.7.post1

AI-powered SRT subtitle generator with Whisper and OpenAI translation

Integrated SRT Subtitle Generator

A tool that uses Whisper speech recognition and OpenAI GPT to generate Korean SRT subtitles.

⚠️ Requirements (read first)

• Python >= 3.10
• OpenAI API key set to OPENAI_API_KEY
• FFmpeg installed and available on PATH (for transcription/whisper features)
• Optional: install extras for whisper/torch

FFmpeg installation (per OS)

• Windows (one of)
  • Chocolatey: choco install ffmpeg
  • Scoop: scoop install ffmpeg
  • Winget: winget install --id=Gyan.FFmpeg -e
• macOS: brew install ffmpeg
• Ubuntu/Debian: sudo apt update && sudo apt install -y ffmpeg
• Fedora: sudo dnf install -y ffmpeg
• Arch: sudo pacman -S ffmpeg

Verify: ffmpeg -version should print version info.

🚀 Features

• Integrated workflow: Speech-to-text → Korean translation → SRT generation
• Flexible execution: Run full process or individual steps
• User-friendly GUI: Intuitive interface
• Simplified structure: Minimal dependencies

📁 Project Structure

SRT_Generator/
├── src/                           # Integrated GUI
│   ├── integrated_srt_generator.py
│   ├── requirements.txt
│   └── run_integrated.bat
├── translator/                    # Translation tools
│   └── src/local_whisper_korean_subtitle_generator/
│       └── tools/
│           ├── korean_translation_tool.py
│           └── srt_formatter_tool.py
├── input/                         # Input files
├── output/                        # Output files
└── README.md

🛠️ Installation and Run

Install from PyPI

pip install ch-srtgen

1. Run the integrated GUI (recommended)

cd src
run_integrated.bat

2. Run individual steps (from source)

# Transcription only
cd src
python integrated_srt_generator.py

# Translation only (when JSON file already exists)
cd src
python integrated_srt_generator.py

📋 Usage

Using the integrated GUI

1. Settings

   • Select Whisper model (tiny ~ large-v3)
   • Select language (auto, ko, en, ja, etc.)
   • Enter OpenAI API key

2. File selection

   • Input file: target video/audio file
   • Output folder: where to save result files

3. Actions

   • Start full process: Transcription → Translation → SRT generation
   • Transcription only: Generate JSON file
   • Translation only: Convert existing JSON to Korean SRT

Workflow

1. Transcription (Whisper)

   • Video/Audio → JSON (timestamp + text)

2. Korean translation (OpenAI GPT)

   • English text → Korean translation

3. SRT generation (Formatter)

   • Translated text → SRT subtitle file

🔧 Configuration

API key setup

Method 1: Environment file (.env) [recommended]

1. Copy env.template to .env
2. Put your real API key in .env:

   OPENAI_API_KEY=your_actual_api_key_here
   

Method 2: Enter via GUI

• Enter the API key in the GUI after launching the app

Method 3: System environment variable

# Windows
set OPENAI_API_KEY=your_actual_api_key_here

# Linux/Mac
export OPENAI_API_KEY=your_actual_api_key_here

⚠️ Security notes:

• Never commit .env to Git
• Do not hardcode API keys in code
• Do not upload API keys to public repos

Model selection

• tiny: Fastest, lower accuracy
• base: Balanced (recommended)
• small: Good accuracy
• medium: Higher accuracy
• large: Highest accuracy, slowest

📝 Output files

• JSON file: filename.json (Whisper result)
• SRT file: filename.srt (final subtitles)

🐛 Troubleshooting

Common issues

1. FFmpeg error: Install FFmpeg
2. API key error: Check OpenAI API key
3. Out of memory: Use a smaller Whisper model

Logs

You can check detailed progress and errors in the GUI log panel.

📄 License

This project is distributed under the MIT License.