Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
Version History
Added safety and confidence guardrails with explicit untrusted-content handling and confidence rubric.
What This Skill Does
Extracts structured cooking recipes from video URLs shared as `recipe <url>`. Uses browser automation to read captions, descriptions, and comments in priority order, falling back to web search and fetch when direct extraction is blocked. Returns a normalized recipe with ingredients, numbered steps, optional metadata, a confidence rating, and source attribution.
Checks captions, pinned comments, and top comments before falling back to search, recovering partial recipes that single-source scrapers miss.
When to use it
- Saving a TikTok cooking video recipe as plain text
- Getting ingredient quantities from an Instagram Reel
- Extracting a YouTube recipe without watching the full video
- Capturing a recipe a friend shared via link in chat
- Archiving a recipe before the original video is deleted
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: recipe https://www.tiktok.com/@chefexample/video/123456789
- 1Opens the URL in browser automation and expands any hidden description text
- 2Captures the full title and caption as the primary source
- 3Loads comments and extracts the pinned or creator comment if present
- 4Scans top comments for ingredient quantities and cooking step patterns
- 5Structures results into a bullet-point ingredient list and numbered steps with source labels
Structured recipe with dish name, ingredients, numbered steps, optional prep time and servings, source notes (Description / Pinned comment / Top comments / Fallback page), and a confidence rating of High, Medium, or Low