webstractor.com

Use extract_public_url when a user names a known public URL or domain and asks to read, extract, summarize, analyze, or retrieve its current content. It supports Hacker News stories, profiles, listings, and story-search URLs; exact public X profile and post URLs; public Wikipedia article URLs; and exact public WeChat Official Account article URLs as well as other supported public pages. Pass a requested page topic, such as pricing, in focus. Use search_web for public webpage discovery and pass site when discovery must stay on one hostname. Use get_x_profile only when the user supplies an exact X handle. Use search_news for current coverage, search_images for openly licensed images, and search_videos for public videos; for a named creator’s latest upload, pass creator, sort=date, and limit=1. Use search_places for named places, addresses, businesses, landmarks, and local categories; pass optional coordinates for nearby-result bias. Use search_stocks when a company or crypto asset name must be resolved to a market symbol, selecting the matching instrument family. Use get_market_movers for daily gainers, losers, and most-active equity rankings. Use get_market_data for current market snapshots with bounded history. When a user requests a currency such as EUR, pass it as quote to get_market_data; do not make a second currency-pair call. get_market_data returns typed schema-v1 data and offers an inline price chart in MCP Apps-capable chats. Extract promising webpage result URLs when full content is needed. Prefer Markdown for reading and JSON for typed fields. Never send credentials or private URLs, and do not use these tools to bypass access controls.