Tavily MCP Server

✓ verified

Web search and extraction tuned for AI agents.

Don't have ToolPiper? We'll take you to the download.

Open repository →

Publisher

Tavily

License

MIT

Transport

Local subprocess

Tools

4

What this server does

Tavily is a search API built for agents: results come back ranked and ready to feed a model, and the extract tool pulls clean content from URLs without the boilerplate. Beyond search it can crawl a site and map its structure, which is useful for grounding answers in a specific documentation set. The free tier covers a meaningful number of monthly calls, and each tool call costs one credit against your plan.

Tools included

  • tavily_search

    Search the web and return ranked results.

  • tavily_extract

    Extract clean content from one or more URLs.

  • tavily_crawl

    Crawl a site starting from a URL.

  • tavily_map

    Map the structure of a website.

The authoritative list comes from the server's tools/list at runtime. This catalog entry may lag the upstream until the next ToolPiper release.

Install in ToolPiper

ToolPiper handles the JSON configuration. Click Install, fill in any required tokens, and the server is ready in your chat.

Don't have ToolPiper? We'll take you to the download.

Install in another client

Install in another MCP client (Claude Desktop, Cursor, Continue, Cline)

Add this to ~/Library/Application Support/Claude/claude_desktop_config.json.

{
  "mcpServers": {
    "tavily": {
      "command": "npx",
      "args": [
        "-y",
        "tavily-mcp"
      ]
    }
  }
}

Example prompts

  • "Search for the latest changes in the MCP specification and summarize them."

  • "Extract the main content from these three blog post URLs."

  • "Crawl docs.example.com and list every page under /guides."

Security notes

The API key gates billing. Treat it as a credential and rotate it from the Tavily dashboard if it leaks. Each tool call counts against your monthly quota. Content returned from the open web is untrusted input; do not let the model act on instructions found inside fetched pages.

Related tools