Get startedPick your language or tool to open the step-by-step guide.
I'll use Tavily with
One API, four primitives
Everything an agent needs to read the web.
Search
One call returns ranked, LLM-ready results with sources and an optional AI answer.Extract
Pull clean, parsed content from any URL — no scraping boilerplate.Crawl
Traverse a site from a starting URL and collect pages in structured form.Map
Discover a site's link graph to plan retrieval and coverage.Resources
Build with Tavily.
Docs
Documentation
Quickstart, concepts, and guides for Search, Extract, Crawl, and Map.Read on Tavily docs ↗Docs
API Reference
Every endpoint and parameter, with request/response examples.Read on Tavily docs ↗GitHub
Python & JS/TS SDKs
Official open-source SDKs: tavily-python and @tavily/core.View on GitHub ↗Docs
MCP Server
Drop Tavily web search into Claude, Cursor, and any MCP client.Read on Tavily docs ↗GitHub
Examples & Notebooks
Reference apps and notebooks: RAG, research agents, and more.Read on Tavily docs ↗Docs