Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
What This Skill Does
Fetches subtitles from YouTube videos over 1 hour using the DownSub API, then translates the English transcript to Chinese verbatim. Handles context limits by processing in 500-line chunks via sub-agents, and outputs a markdown file with an executive summary and a key metrics table.
Splits work across sub-agents in 500-line chunks, so videos that exceed a single session's context limit can still be fully transcribed and translated.
When to use it
- Transcribing a 2-hour conference keynote for offline reading
- Translating an English YouTube lecture into Chinese study notes
- Extracting a full verbatim record of a long YouTube interview
- Summarizing a multi-hour documentary into key talking points
- Converting a lengthy coding tutorial into a searchable markdown document
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: Translate this 90-minute YouTube video to Chinese
- 1Call DownSub API with the video URL to retrieve available subtitle tracks
- 2Verify the returned language is 'en' or 'en-auto', stop if another language is detected
- 3Spawn a sub-agent to translate the transcript in 500-line chunks, writing each chunk to a separate file
- 4Sub-agent merges chunks into full_transcript.md and prepends a Chinese executive summary and key metrics table
- 5Return the file path to the user, or upload via zhiyan MCP if the tool is available
full_transcript.md containing the complete Chinese translation, an executive summary, and a key metrics table
Requirements
Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.