good-TTvideo2text

good-TTvideo2text

Extract audio from Douyin/TikTok videos and transcribe to text using ASR service.

Task Objective

Convert Douyin/TikTok video audio to text with timestamps, supporting both interactive UI and command-line workflow.

Capabilities: Video parsing, audio extraction, ASR transcription, timestamp generation

Trigger: User provides Douyin/TikTok URL and requests transcription

Usage Modes

Mode 1: Web UI (Recommended)

Visual interface for transcription management:

# Install dependencies
cd skills/good-TTvideo2text
pip install -r requirements.txt

# Start service (default port 8000)
python app/main.py

# Browser access
http://localhost:8000

Features:

• Paste video URL for instant transcription
• View results with timestamps
• Cookie management for restricted videos
• Real-time progress updates

Mode 2: Script (Command Line)

Suitable for automation, AI workflow integration:

# Basic usage
python scripts/transcribe.py "https://v.douyin.com/xxx"

# Extract URL from share text
python scripts/transcribe.py "7.47 复制打开抖音，看看【用户名】的作品 https://v.douyin.com/xxx"

# JSON output
python scripts/transcribe.py "https://v.douyin.com/xxx" --output json

# Text output (default)
python scripts/transcribe.py "https://v.douyin.com/xxx" --output text

Output Format (Text):

=== Video Info ===
Title: Video title
Author: Author name
Duration: 30s

=== Transcription ===
Full Text:
Complete transcription text...

Sentences with Timestamps:
[00:00-00:03] First sentence
[00:03-00:06] Second sentence

Output Format (JSON):

{
  "success": true,
  "video_info": {
    "title": "Video title",
    "author": "Author name",
    "duration": 30,
    "create_time": 1234567890
  },
  "transcription": {
    "text": "Complete transcription...",
    "sentences": [
      {
        "start_ms": 0,
        "end_ms": 3000,
        "text": "First sentence"
      }
    ]
  }
}

Prerequisites

Dependencies:

• fastapi>=0.120.3, uvicorn>=0.35.0, httpx>=0.28.1
• TikTokDownloader dependencies (see requirements.txt)