Tavily MCP Server
Official Tavily MCP server: real-time web search, page content extraction, site mapping and crawling
- 2,462 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Sends queries and page URLs to an external service
- Requires a secret API key
Install
- Works in
Manual install
claude mcp add --transport http tavily https://mcp.tavily.com/mcpAfter adding, open /mcp and complete OAuth. The key can also be passed in the URL: ?tavilyApiKey=<key>.
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Sends queries and page URLs to an external service
- Requires a secret API key
This is third-party code. Review the repository files before installing.
What it does
The server gives the agent four tools on top of the Tavily API: search, page content extraction, site structure mapping and link-following crawl. It connects to Tavily's hosted server over HTTP or runs locally via npx. Search defaults such as depth and result count can be preset. Authentication uses an API key or OAuth.
Who it is for. For developers and researchers who need the agent to search and read current web data.
Good fit when
- The agent needs fresh web data rather than model knowledge
- You need to pull text from several pages for analysis
- You need to crawl documentation or a site and collect its structure
Not a fit when
- The agent already has built-in web search that is enough
- Queries must not leave your infrastructure
Example request
Find recent news about the latest Node.js release and summarize the changes with linksMCP
- Transport
- stdio, http
- Authentication
- API key
Environment variables
TAVILY_API_KEYrequired, secret- Tavily API key for local mode
DEFAULT_PARAMETERS- JSON with default search parameters
TAVILY_HUMAN_ID- End user identifier for analytics
Limitations
Requires a Tavily account and API key; the free tier is limited. Local mode needs Node.js 20 or later. All queries go to Tavily servers; availability from Russia without a VPN is unverified.
How to disable. Run claude mcp remove tavily or delete the tavily-mcp block from your MCP client config.
FAQ
Remote or local server?
Remote is simpler: nothing to install and OAuth is supported. Local via npx is for clients without HTTP transport.
Can search parameters be preset?
Yes, with the DEFAULT_PARAMETERS variable locally or the DEFAULT_PARAMETERS header on the remote server.
Discussion
No questions or reviews yet.
Related
MCP serversA unified search MCP server: one interface to Tavily, Brave, Kagi, Exa, Linkup, Firecrawl and GitHub search
MCP server for real-time web search: general and vertical search, batch queries and page content extraction
A web search MCP server that gives the agent full page text at once: Stack Overflow answers, GitHub threads, arXiv papers
Official Firecrawl MCP server: web search, page scraping to Markdown, site crawling and in-page actions
Web search via DuckDuckGo and page reading for the agent with no API keys, with built-in rate limiting
Yandex and Google search MCP server
web-compass
An MCP web-search and content-extraction server via Yandex and Google on top of SerpAPI, geared toward RU/CIS results
Comparison with similar tools
Entries that share the main search query, and how they differ.
| Entry | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| Tavily MCP Server (this entry) | yes | API key | 2,462 | |
| AnySearch Skill | yes | 6,797 | a web search skill with no MCP server | |
| Tavily Agent Skills | yes | 491 | eight Tavily skills built on the tvly CLI | |
| fastCRW | yes | not required | 1,192 | a Rust server that turns pages into Markdown or JSON |
README in short
The README covers connecting the hosted server to Claude Code and Cursor with the key in the URL or via OAuth, plus local launch through npx. It explains search defaults and the optional TAVILY_HUMAN_ID variable for per-user analytics. Requirements are a Tavily key and Node.js 20. MIT licensed.