English | 简体中文

VideoHub is a desktop video workflow application built with PyQt6. It combines multi-platform media processing, speech transcription, bilingual subtitle generation, LLM summaries, idle-time queue scheduling, browser extension integration, and live recording utilities in a single tool.

It is designed for users who want to turn online or local media into reusable assets: local video/audio files, transcripts, subtitles, and structured markdown summaries.

This project is intended for lawful and authorized use only. Please read DISCLAIMER.md before using features related to third-party platform content, cookies, tokens, downloading, or recording.

• The live recording feature is currently not fully functional if the project root does not contain live_recorder/.
• src/douyin/downloader.py still reports that Douyin user-profile batch download is not implemented.
• Some live platform detection branches in src/live_recorder_adapter.py are incomplete.
• Parts of the application are still centered around a large single-file GUI controller in main.py.

• Python 3.8+
• Windows is the primary tested platform
• FFmpeg for subtitle/video processing and live recording
• Optional: Chrome/Edge for the browser extension
• Optional: CUDA-capable GPU for faster Whisper transcription

git clone [email protected]:cacity/VideoHub.git
cd VideoHub

# Optional but recommended
conda create -n VideoHub python=3.12
conda activate VideoHub

pip install -r requirements.txt

python main.py

This launches the PyQt desktop GUI and also starts the local Flask API server on port 8765 for queue integration.

python src/youtube_transcriber.py --help
python src/douyin_cli.py "https://v.douyin.com/xxxxx/"
python src/ffmpeg_config_cli.py help

Create a .env file in the project root if you want summary generation or custom API endpoints.

OPENAI_API_KEY=your_openai_key
OPENAI_BASE_URL=
OPENAI_MODEL=
DEEPSEEK_API_KEY=
KOUSHARE_ACCESS_TOKEN=

You can also configure many options from the GUI settings page instead of editing .env manually.

Run the main application:

python main.py

Main tabs in the GUI include:

• Online Video
• Local Audio / Video
• Batch Processing
• Idle Queue
• Live Recorder
• Processing History
• Settings

src/youtube_transcriber.py is the main reusable CLI for transcription, subtitles, summaries, template management, batch jobs, and cleanup.

# Process a single YouTube video
python src/youtube_transcriber.py --youtube "https://www.youtube.com/watch?v=VIDEO_ID"

# Process video and generate subtitles
python src/youtube_transcriber.py --youtube "https://www.youtube.com/watch?v=VIDEO_ID" --download-video --generate-subtitles

# Burn subtitles into the processed/local video
python src/youtube_transcriber.py --youtube "https://www.youtube.com/watch?v=VIDEO_ID" --download-video --generate-subtitles --embed-subtitles

# Process a local audio file
python src/youtube_transcriber.py --audio "path/to/file.mp3"

# Process a local video file
python src/youtube_transcriber.py --video "path/to/file.mp4" --generate-subtitles

# Generate a summary directly from text
python src/youtube_transcriber.py --text "path/to/file.txt"

# Process multiple URLs
python src/youtube_transcriber.py --urls "<url1>" "<url2>"

# Preview cleanup
python src/youtube_transcriber.py --cleanup-preview

For Douyin single-link processing, use the dedicated CLI:

python src/douyin_cli.py "https://v.douyin.com/xxxxx/"
python src/douyin_cli.py "https://www.douyin.com/video/xxxxx" -o "workspace/douyin_downloads"

Notes:

• It is designed for single-video processing workflows.
• It expects the required Douyin backend/service to be available.
• User-profile batch download is not implemented in the current backend.

VideoHub includes dedicated Koushare support.

Current behavior:

• Login/token logic lives in the GUI and in src/koushare_downloader.py
• Authorization uses the raw token value, not Bearer <token>
• Higher-quality playback (such as FHD) may require a valid login token

Recommended usage:

• Open the desktop app
• Go to the settings tab
• Log in with your Koushare account or paste an existing token
• Then process the target Koushare URL through the standard workflow

The chrome_extension/ folder contains the local browser extension.

Typical flow:

1. Load the unpacked extension in Chrome/Edge
2. Start python main.py
3. Open a supported video page
4. Click the injected button to add the task into the local idle queue

The extension communicates with the desktop app through the local API server.

Once the GUI is running, the local API is available on http://127.0.0.1:8765.

• GET /api/health
• GET /api/queue
• POST /api/queue/add
• DELETE /api/queue/clear
• DELETE /api/queue/remove/<task_id>
• GET /api/settings
• PUT /api/settings

curl http://127.0.0.1:8765/api/health
curl http://127.0.0.1:8765/api/queue
curl -X POST http://127.0.0.1:8765/api/queue/add -H "Content-Type: application/json" -d '{"platform":"youtube","url":"https://example.com","title":"sample"}'

Required fields for queue insertion:

• platform
• url
• title

VideoHub ships with an FFmpeg management CLI:

python src/ffmpeg_config_cli.py status
python src/ffmpeg_config_cli.py test
python src/ffmpeg_config_cli.py mode auto
python src/ffmpeg_config_cli.py path "C:/ffmpeg/bin/ffmpeg.exe"
python src/ffmpeg_config_cli.py download

Use it to inspect current FFmpeg status, switch execution modes, set a custom binary path, or install/configure FFmpeg.

Runtime-generated files are placed under workspace/.

workspace/
  videos/
  downloads/
  subtitles/
  transcripts/
  summaries/
  videos_with_subtitles/
  native_subtitles/
  douyin_downloads/
  twitter_downloads/
  bilibili_downloads/
  live_downloads/

VideoHub/
├── main.py
├── src/
│   ├── youtube_transcriber.py
│   ├── douyin_cli.py
│   ├── ffmpeg_config_cli.py
│   ├── api_server.py
│   ├── koushare_downloader.py
│   ├── live_recorder_adapter.py
│   └── subtitle_merger.py
├── chrome_extension/
├── workspace/
├── templates/
├── logs/
├── idle_queue.json
└── README.md

• Paste the playlist/video URL into the GUI
• Enable subtitle generation
• Optionally translate subtitles
• Save outputs under workspace/

• Start python main.py
• Add items from the GUI or browser extension
• Let the idle queue run during the configured time window

• Copy the Douyin link or share text
• Extract the URL if needed
• Run python src/douyin_cli.py "<douyin-url>"

• Run python src/youtube_transcriber.py --video "path/to/file.mp4"
• Configure API keys if you want LLM summary generation
• Review transcript, subtitles, and summary in workspace/

There is no unified automated test suite documented in the repository yet. For smoke testing, use the existing entry points:

python src/youtube_transcriber.py --help
python src/douyin_cli.py --help
python src/ffmpeg_config_cli.py help
python main.py

This project is licensed under the MIT License.

• PyQt6
• yt-dlp
• OpenAI Whisper
• Flask

If this project helps you, a star is welcome.