> ## Documentation Index
> Fetch the complete documentation index at: https://docs.portkey.ai/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Tavily MCP server

> The Tavily MCP server enables AI agents to perform real-time web search, extraction, crawling, mapping, and deep research through MCP.

## When you should use this server

* Search the web in real time for current information.
* Extract clean content from one or more URLs.
* Crawl documentation sites, blogs, or domains to gather multiple pages.
* Map a website's structure before deciding what to crawl or extract.
* Run deep research workflows for complex, multi-step queries.
* Add Tavily-powered retrieval and research into MCP-compatible AI clients.

## Key features

* Real-time web search with AI native retrieval for Agents.
* Clean content extraction from web pages
* Site crawling across multiple linked pages
* URL discovery and site mapping
* Deep research for intensive, multi-step queries
* Flexible upstream authentication through API key or OAuth
* Hosted remote MCP support for compatible clients

## Authentication

### Option 1: Tavily API key

* **Method:** API key (`TAVILY_API_KEY`)
* **Use when:** You do not have an MCP-compatible client
* **Notes:** Use the Tavily API key in the hosted MCP URL or through header-based auth

### Option 2: OAuth

* **Method:** OAuth 2.0
* **Use when:** Your MCP client supports OAuth with Tavily's hosted remote MCP server
* **Notes:** OAuth is handled on the Tavily server side

## Endpoints

* **Remote hosted URL:** `https://mcp.tavily.com/mcp/?tavilyApiKey=<your-api-key>`
* **Remote via OAuth:** `https://mcp.tavily.com/mcp/`
* **Local via npx:** `npx -y mcp-remote https://mcp.tavily.com/mcp/?tavilyApiKey=<your-api-key>`
* **npm installation:** `npm i tavily-mcp`

## Setup and usage

### Environment variables

```bash theme={"system"}
TAVILY_API_KEY=tvly-YOUR_API_KEY
```

## Tools provided

### tavily\_search

Search the web in real time with filters for domains, dates, and search depth.

### tavily\_extract

Extract clean markdown or text content from one or more web pages.

### tavily\_crawl

Crawl websites across linked pages with configurable depth, breadth, and path selection.

### tavily\_map

Discover a website's structure and list relevant URLs before extracting content.

### tavily\_research

Run deep research workflows that combine search, extraction, and synthesis for complex questions.

## Notes

* Tavily supports both API key and OAuth-based authentication
* Hosted remote MCP is the simplest way to use Tavily MCP