Tavily MCP

Search, extract, map and crawl live web content through Tavily's official hosted MCP server.

Quick verdict

A focused, well-documented web-research profile that sits between simple search and full browser automation, with the main risks concentrated in data trust, privacy and credit usage.

MaintainerTavily (official)
Security scopeRead-only
Transportsstreamable-http, stdio
Install channelremote · https://mcp.tavily.com/mcp
EvidenceSource Validated
LicenseMIT · Open source

Tavily MCP gives assistants a first-party hosted interface for real-time web search, clean page extraction, site mapping and bounded crawling. The proposed configuration uses a bearer header rather than placing the Tavily API key in the URL, and the profile remains read-oriented even when a crawl spans many pages.

MCP security best practices

Install Tavily MCP

http https://mcp.tavily.com/mcp

Config entry (mcpServers)

{
  "mcpServers": {
    "tavily": {
      "url": "https://mcp.tavily.com/mcp",
      "headers": {
        "Authorization": "Bearer <TAVILY_API_KEY>"
      }
    }
  }
}

Tavily-hosted endpoint. OAuth is also supported; an API key in the query string is documented but deliberately not used here — the bearer header keeps the endpoint URL secret-free. The MIT-licensed npm server can instead run locally over stdio, still calling Tavily's hosted API.

Credentials

Name	Required	What it is for
`TAVILY_API_KEY`	Required	Tavily API key sent through the MCP client's secure Authorization header. Use a dedicated key with account-level usage controls where available.

Verified configurations

Not run by us yet. This profile is Source Validated: the install channel, config entry, credentials and license were checked against the source and upstream metadata, but no configuration has been installed, discovered and executed in a recorded environment. When one is, it appears here with the eight scope dimensions below.

Best for

Agent and RAG workflows that need current web results plus structured extraction or controlled site-wide discovery from one API.

Why use it

It combines four complementary retrieval tools behind one official endpoint and returns content shaped for AI workflows without requiring local scraping infrastructure.

Tradeoffs

Search and crawl results are untrusted external input, broad site operations can consume credits and time, and the hosted service receives queries and target URLs. Use domain and result limits, protect the API key, preserve returned sources, and avoid open-ended crawls when a single extract will do.

Upstream state

repository active · in the official MCP Registry (active). Lifecycle is derived from a committed upstream snapshot at build time, never authored by hand.

Official MCP Registry name: io.github.tavily-ai/tavily-mcp

Editorial review

Reviewed byOpenSourcesAI EditorialLast updatedAugust 2026SourcesOfficial Tavily MCP documentation; Official Tavily MCP source repository; Tavily pricing FAQ; Tavily MCP license; Official MCP Registry — Tavily

Grade B — Source Validated. Re-review is due by 2027-02-18. Config entries, credential names and package status can change upstream; verify against the source repository before granting an agent access.

Sources

Official Tavily MCP documentation (tier 1, checked 2026-08-19)
Official Tavily MCP source repository (tier 1, checked 2026-08-19)
Tavily pricing FAQ (tier 1, checked 2026-08-19)
Tavily MCP license (tier 1, checked 2026-08-19)
Official MCP Registry — Tavily (tier 2, checked 2026-08-19)

AWS Documentation MCPSearch and read official AWS documentation through MCP without giving the agent credentials to your AWS account.
Exa MCPSearch the web and fetch clean page content through Exa, with optional advanced search and multi-step research tools.
FetchReference MCP server that fetches a web page and returns it as Markdown, so a local model can read live content without a browser.

Related guides

All MCP servers →