📝

OpenClaw YouTube Transcript

Transcribe YouTube videos to text by extracting captions and subtitles directly from the video URL using yt-dlp, without any audio processing or local models.

by Yoav

Productivity & TasksProductivityOpen SourceCLI

# OpenClaw YouTube Transcript Transcribe YouTube videos to text by extracting captions and subtitles directly from the video URL using yt-dlp, without any audio processing or local models. This skill extracts transcripts from YouTube videos by fetching captions and subtitles directly using yt-dlp. It converts any YouTube video to readable text without downloading audio or running speech-to-text models, making it fast and lightweight. Made specifically for OpenClaw agents. ## How It Works The skill uses a Python script that calls yt-dlp to fetch subtitle data (both manual captions and YouTube auto-generated subtitles) from a given video URL. The transcript is either printed to stdout or saved to a file. No audio processing or local AI models are involved — it works entirely by extracting existing caption data from YouTube. ## Key Features - **Direct Caption Extraction**: Fetches subtitles directly from YouTube without downloading or processing audio - **Auto-Generated Subtitle Support**: Works with both manually uploaded captions and YouTube's auto-generated subtitles - **Multi-Language Support**: Specify subtitle language code (en, es, fr, etc.) to get transcripts in different languages - **File Output**: Optionally save transcripts to a text file instead of printing to stdout - **Lightweight**: No audio models or heavy dependencies, just yt-dlp and Python ## Requirements - **yt-dlp**: Command-line tool for extracting video data from YouTube. Install via brew install yt-dlp (macOS) or apt install yt-dlp (Linux) - **Python 3**: Required to run the transcription script ## Use Cases - **Video Summarization**: Transcribe a video first, then have the assistant summarize or analyze the content - **Content Research**: Extract spoken content from educational or informational videos for reference - **Accessibility**: Get text versions of videos for reading instead of watching - **Quote Extraction**: Pull specific quotes or statements from video content ## Installation Install via: `npx clawhub@latest install openclaw-youtube-transcript`