youtube-transcriber

What This Skill Does

The YouTube Transcriber skill provides a robust, one-command solution for converting any YouTube video into text. Whether you need to extract information from a lecture, summarize a long-form video, or simply have a reference copy of spoken dialogue, this tool streamlines the process. It is highly intelligent: it first checks for native YouTube subtitles to save you time and cost. If those are missing or disabled, it seamlessly pivots to downloading the audio and processing it through the OpenAI Whisper API to generate an accurate transcript. This ensures you get high-quality results for 99+ languages, regardless of the video's original accessibility settings.

Installation

To get started, ensure you have the necessary system-level dependencies installed. The script relies on yt-dlp for video/audio extraction and ffmpeg for audio conversion, both of which can be installed via Homebrew (brew install yt-dlp ffmpeg) or pip/package managers. You must also have your OPENAI_API_KEY exported as an environment variable in your terminal session or shell configuration file. Once prerequisites are met, install the skill via the OpenClaw hub: clawhub install openclaw/skills/skills/edisonchenai/youtube-transcriber.

Use Cases

• Academic Research: Extracting transcripts from educational videos or webinars for study notes.
• Content Creation: Repurposing video content into blog posts or newsletters.
• Accessibility: Creating text transcripts for hearing-impaired users or for SEO indexing purposes.
• Efficiency: Quickly searching through lengthy video content by keywords without watching the entire video.

Example Prompts

1. "Transcribe this video for me: https://www.youtube.com/watch?v=dQw4w9WgXcQ"
2. "Can you give me a full transcript of the video I just sent, and please save it to my notes folder?"
3. "What does the speaker say in this clip? Please provide a high-accuracy transcript using the Whisper API."

Tips & Limitations

• Cost Efficiency: Always leverage native subtitles when possible, as they are free. Use the --force-whisper flag only when high accuracy is required or native captions are insufficient.
• API Limits: Keep in mind that Whisper API usage costs roughly $0.006 per minute. Long videos may take slightly longer to process as audio is compressed to fit the 25MB upload constraint.
• Maintenance: If you receive a 403 error, run pip install -U yt-dlp immediately to update your extraction engine, as YouTube frequently updates their anti-scraping measures.