Official MCP server that gives AI agents and MCP clients web search, scraping, crawling, browser interaction and Markdown or structured-data extraction via Firecrawl.
tavily-ai/tavily-mcp
An MCP server giving AI assistants real-time web search, content extraction, site mapping and crawling through Tavily's API.
- Stars
- 2,452
- Forks
- 303
- Open issues
- 63
- Last commit
- 21 hours ago
What is it?
Tavily MCP is a Model Context Protocol server that exposes Tavily's web tools to MCP-compatible clients such as Claude Desktop, Claude Code and Cursor. It provides four tools: tavily-search for real-time web search, tavily-extract for pulling content out of web pages, map for building a structured map of a website, and crawl for systematically exploring a site.
The project can run as a local process through npx (tavily-mcp@latest, Node.js 20 or higher), or connect to Tavily's hosted remote MCP endpoint over HTTP. The remote option authenticates with an API key in the URL or Authorization header, or through an OAuth flow. Default search parameters can be set with a DEFAULT_PARAMETERS setting, and an optional TAVILY_HUMAN_ID value can be forwarded for per-user analytics.
Key features
- 01Real-time web search through the tavily-search tool.
- 02Content extraction from web pages via tavily-extract.
- 03Structured site mapping and systematic website crawling.
- 04Remote hosted endpoint with API key or OAuth authentication.
- 05Local run via npx with configurable default search parameters.
Use cases
- Giving a coding assistant or chat model access to current web information while it works.
- Building a knowledge graph by combining web extraction with a graph database MCP server.
- Exploring a website's structure before scraping or auditing its pages.
- Integrating web research into Cursor or VS Code workflows via Cline.
Quick start
- 1
Get a Tavily API key from tavily.com
- 2
Run the server locally with NPX
npx -y tavily-mcp@latest - 3
Or connect to the remote MCP server with Claude Code
claude mcp add --transport http tavily https://mcp.tavily.com/mcp/?tavilyApiKey=<your-api-key>
Things to know
- Using the tools requires a Tavily API key, which needs a free or paid account.
Topics
FAQ
- What is tavily-mcp?
- An MCP server giving AI assistants real-time web search, content extraction, site mapping and crawling through Tavily's API. Tavily MCP is a Model Context Protocol server that exposes Tavily's web tools to MCP-compatible clients such as Claude Desktop, Claude Code and Cursor. It provides four tools: tavily-search for real-time web search, tavily-extract for pulling content out of web pages, map for building a structured map of a website, and crawl for systematically exploring a site.
- Is tavily-mcp free to use? What license does it have?
- tavily-mcp is open source under the MIT license. Check the license terms before using it in commercial products.
- Is tavily-mcp still maintained?
- Actively maintained. The last commit to tavily-mcp was on Oct 21, 2026.
- What can I use tavily-mcp for?
- Giving a coding assistant or chat model access to current web information while it works. Building a knowledge graph by combining web extraction with a graph database MCP server. Exploring a website's structure before scraping or auditing its pages. Integrating web research into Cursor or VS Code workflows via Cline.
- How do I install tavily-mcp?
- Get a Tavily API key from tavily.com Run the server locally with NPX: npx -y tavily-mcp@latest Or connect to the remote MCP server with Claude Code: claude mcp add --transport http tavily https://mcp.tavily.com/mcp/?tavilyApiKey=<your-api-key>
- Where can I get tavily-mcp?
- The source code and installation instructions are on GitHub: https://github.com/tavily-ai/tavily-mcp
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
Local-first web search, fetch, crawl and research tools for AI agents over MCP, with no API keys and no per-query cost.
A Python metasearch library that queries multiple web search backends and returns unified text, image, video, news and book results.
An MCP server giving AI agents real-time web search, vertical domain search, parallel batch search and URL-to-Markdown page extraction.
A free, keyless web search, X search and page fetch plugin that gives DeepSeek and other text-only models live web access.
An MCP server that connects AI agents to Exa for web search, page content fetching, and multi-step research.