Tavily Extract Skill

Extract clean, readable content from one or more specific URLs. Returns markdown-formatted text, stripping navigation, ads, and boilerplate. Handles JavaScript-rendered pages with advanced mode.

Authentication

Set TAVILY_API_KEY in ~/.openclaw/.env.local:

TAVILY_API_KEY=tvly-your-api-key-here

When to Use This Skill

• You have specific URLs and need their content (e.g. from a previous search result)
• Reading documentation pages, blog posts, or articles at known URLs
• Extracting content from JavaScript-rendered pages (SPAs, React apps)
• Getting only the relevant portion of a page (via query + chunks_per_source)