Video2Text

A professional tool for automatically extracting text content from Douyin videos, supporting both CLI and Web interfaces.

中文文档

Features

• Auto-download Douyin videos
• AI-powered speech recognition transcription
• Multiple output formats: TXT, SRT, VTT, Markdown
• CLI and Web interface
• Batch processing support
• Local processing for privacy protection

System Requirements

• Node.js: >= 18.0.0
• yt-dlp: Video download tool
• ffmpeg: Audio/video processing
• f2 (optional): Optimized for Douyin downloads

Installation

1. Install System Dependencies

macOS:

brew install yt-dlp ffmpeg
brew tap fyrfyrr/f2 && brew install f2  # optional

Linux:

pip install yt-dlp
sudo apt install ffmpeg

Windows:

choco install yt-dlp ffmpeg

2. Install Project Dependencies

git clone https://github.com/wangjs-jacky/video2text.git
cd video2text
npm install

3. Download Whisper Model

cd node_modules/whisper-node/lib/whisper.cpp/models
bash download-ggml-model.sh base
cd -

Usage

CLI

# Basic usage
npm run cli extract https://v.douyin.com/xxx/

# Specify format and output
npm run cli extract https://v.douyin.com/xxx/ -f srt -o ./subs

# Batch processing
npm run cli extract --file links.txt

# Start web server
npm run cli serve --port 3000

API

# Extract video
curl -X POST http://localhost:3000/api/extract \
  -H "Content-Type: application/json" \
  -d '{"url": "https://v.douyin.com/xxx/", "format": "txt"}'

Options

| Parameter | Default | Description | |-----------|---------|-------------| | --format | txt | Output format (txt/srt/vtt/md) | | --output | ./output | Output directory | | --model | base | Whisper model (tiny/base/small/medium/large-v3) | | --cookie | - | Douyin cookie for login-required videos |

Tech Stack

• Node.js >= 18, TypeScript
• Express (Web), CAC (CLI)
• ffmpeg, yt-dlp, f2
• whisper.cpp for AI transcription

License

MIT