Tavily MCP

Tavily's own MCP server gives your agent real-time access to the web through Tavily's API. It can search for current sources, pull clean content from pages, map or crawl a whole site, and run multi-source research that comes back as a detailed answer. What you can do with it - Search the web for current information, news and sources, with filters for dates, topic and domains - Extract the main content from one or more URLs - Map a site's structure, or crawl many pages across it - Run multi-source research on a topic and get a detailed answer How it connects This listing runs the server on your machine with npx -y tavily-mcp@latest (Node.js 20 or later), with your Tavily API key in TAVILY_API_KEY. You can get a key with a free Tavily account. Tavily also hosts the same tools at https://mcp.tavily.com/mcp/, where you sign in to Tavily with OAuth or pass your API key; Tavily documents that setup for Claude, ChatGPT, Codex, Cursor, VS Code and more. Good to know - Set DEFAULT_PARAMETERS to a JSON object (for example max_results, search_depth or include_images) to change the defaults for every search. - Requests use your Tavily account's credits and limits.

• “Can you search for recent developments in quantum computing?”

  1. tavily_search
  Copy

• “Search for news articles about AI startups from the last 7 days.”

  1. tavily_search
  Copy

• “Search for climate change research on nature.com and sciencedirect.com.”

  1. tavily_search
  Copy

• “Extract the main content from this article: https://example.com/article”

  1. tavily_extract
  Copy

• “Crawl Stripe's subscription documentation and explain the billing flow.”

  1. tavily_crawl
  Copy

• “Research the major AI chip vendors and compare them, with sources.”

  1. tavily_research
  Copy