Exa MCP server

Endpoint: https://mcp.exa.ai/mcp · Authentication: No auth

Exa web search and page-fetching tools for retrieving current web content. Works without an API key on a free tier; signing in raises limits.

Test it live Run the release-readiness report

How do I authenticate?

Add the endpoint directly in your client. The catalog records no authentication requirement; no OAuth sign-in or API key is configured in this setup. The live server may change.

Authentication and connection guide

Connection specification

Remote endpoint: https://mcp.exa.ai/mcp
Browser access: Not yet measured
Declared MCP transport: Streamable HTTP
Live-validated MCP transport: Streamable HTTP — Observed by the latest catalog validation
Declared authentication: No authentication
Detected authentication: No authentication detected or retained from publisher evidence
Protocol lifecycle: Stateful MCP · 2025-11-25
Authorization server: https://auth.exa.ai
Category: Search

Product documentation · Source repository · Official listing documentation

More Search MCP servers · Browse all MCP servers · Test this endpoint in the MCP Playground

Authorization setup

No app registration needed

This catalog entry does not require authorization or provider app registration.

Connect it

Live-validated endpoint: https://mcp.exa.ai/mcp · Streamable HTTP. No authentication is required by the catalog listing.

Agent prompt for Claude Code or Codex

Add the Exa MCP server to my MCP client configuration and verify that it works.

What it is: Exa web search and page-fetching tools for retrieving current web content. Works without an API key on a free tier; signing in raises limits.
Server URL: https://mcp.exa.ai/mcp
Transport: Streamable HTTP (remote MCP server, no local install).
Authentication: none. The server accepts connections without credentials.
Tools (2, as last observed by mcptest): web_fetch_exa, web_search_exa.
Reference: https://mcptest.io/servers/exa/

Steps:
1. Work out which MCP client I use (Claude Code, Codex, Cursor, VS Code or another) and add the server under the name "exa" in that client's MCP configuration, user scope unless I ask otherwise.
   Claude Code: claude mcp add --transport http --scope user 'exa' 'https://mcp.exa.ai/mcp'
   Codex: codex mcp add 'exa' --url 'https://mcp.exa.ai/mcp'
   Other clients: use the setup for that client on the reference page.
2. Never write a credential value into a file that could be committed; reference an environment variable or the client's secret store, and ask me for any value you need.
3. Verify: connect to the server and call tools/list. Confirm the tools above are listed, and report any sign-in step I still have to complete.

Claude Code setup

Run in a terminal. This user-scoped entry is available across projects.

claude mcp add --transport http --scope user 'exa' 'https://mcp.exa.ai/mcp'

Codex CLI setup

Run in a terminal; Codex stores the entry in ~/.codex/config.toml.

codex mcp add 'exa' --url 'https://mcp.exa.ai/mcp'

Cursor setup

Add to .cursor/mcp.json for this project, or ~/.cursor/mcp.json for all projects.

{
  "mcpServers": {
    "exa": {
      "url": "https://mcp.exa.ai/mcp"
    }
  }
}

VS Code setup

Add to .vscode/mcp.json, or run “MCP: Open User Configuration” for a private user-level entry.

{
  "servers": {
    "exa": {
      "type": "http",
      "url": "https://mcp.exa.ai/mcp"
    }
  }
}

Claude setup

Add the endpoint above in Claude:

Customize > Connectors > Add custom connector
Team/Enterprise: Organization settings > Connectors > Add > Custom > Web

Claude Desktop setup

Add the endpoint above in Claude Desktop:

Remote: Customize > Connectors (same as claude.ai)
Local: Settings > Developer > Edit Config, or Settings > Extensions for .mcpb files
~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows), local servers only

ChatGPT setup

Add the endpoint above in ChatGPT:

ChatGPT web: Plugins > Add custom MCP server, then configure authentication and accept the risk warning

Windsurf setup

Add the endpoint above in Windsurf:

~/.config/devin/mcp_config.json (%APPDATA%\devin\mcp_config.json on Windows), serverUrl or url
Cascade panel > ... > MCPs > Open MCP config file

Gemini CLI setup

Add the endpoint above in Gemini CLI:

gemini mcp add --transport http <name> <url>
~/.gemini/settings.json (user) or .gemini/settings.json (project), mcpServers

MCP SDK setup

Add the endpoint above in MCP SDK:

Programmatic: new Client(...).connect(new StreamableHTTPClientTransport(url))

Client compatibility

Client	Status	Why
Claude Code	Supported	Claude Code supports Streamable HTTP servers and servers without authentication.
Codex	Partial	Codex's documentation does not state support for servers without authentication. Codex CLI ignores resources: The server can still connect, but Codex CLI does not consume the advertised resources capability. Codex CLI ignores prompts: The server can still connect, but Codex CLI does not consume the advertised prompts capability.
Cursor	Supported	Cursor supports Streamable HTTP servers and servers without authentication.
VS Code	Supported	VS Code supports Streamable HTTP servers and servers without authentication.
Claude	Supported	Claude supports Streamable HTTP servers and servers without authentication.
Claude Desktop	Supported	Claude Desktop supports Streamable HTTP servers and servers without authentication.
ChatGPT	Partial	ChatGPT ignores resources: The server can still connect, but ChatGPT does not consume the advertised resources capability. ChatGPT ignores prompts: The server can still connect, but ChatGPT does not consume the advertised prompts capability. Stateful routing needs operational care: ChatGPT can preserve MCP sessions, but the deployment must route a session consistently and retain its state.
Windsurf	Partial	Windsurf's documentation does not state support for servers without authentication.
Gemini CLI	Supported	Gemini CLI supports Streamable HTTP servers and servers without authentication.
MCP SDK	Supported	MCP SDK supports Streamable HTTP servers and servers without authentication.

Computed from this server's validated transport, protocol and authentication against documented client support. Checks the catalog has not observed are not counted.

Similar servers

Jina AI — Jina Reader, web, arXiv, and image search, screenshot, PDF extraction, and reranking tools. Some tools work without a key; an optional Jina API key raises limits and unlocks the rest. No authentication
Tavily — Tavily search, extract, crawl, and map tools for real-time web research. Clients sign in with OAuth, or append a Tavily API key as the tavilyApiKey query parameter. OAuth 2.1
Firecrawl — Firecrawl scrape, crawl, map, search, and structured-extraction tools that turn websites into LLM-ready Markdown or JSON. This endpoint signs in with OAuth; API-key clients use the /v2/mcp endpoint with a bearer token. OAuth 2.1
Grep by Vercel — Searches code across about a million public GitHub repositories by literal string or regular expression, with language, repository, and path filters. No authentication is required. No authentication
Parallel Search — Searches the web and fetches pages as compact markdown through Parallel's Search and Extract APIs. Works anonymously on a free tier; a Parallel API key as a bearer token raises the limits. No authentication
Parallel Task — Starts asynchronous Parallel deep-research and enrichment tasks, polls their status, and returns results as markdown. Supports OAuth or a Parallel API key as a bearer token. OAuth 2.1

Guides for this server

Watch this server for breaking changes

Share and embed

Maintain Exa? Embed this badge in your README. It follows the latest catalog validation result and links to this server profile.

Markdown

[![Tested on mcptest.io](https://mcptest.io/badge/server/exa.svg)](https://mcptest.io/servers/exa/)

HTML

<a href="https://mcptest.io/servers/exa/"><img src="https://mcptest.io/badge/server/exa.svg" alt="Tested on mcptest.io"></a>

reStructuredText

.. image:: https://mcptest.io/badge/server/exa.svg
   :alt: Tested on mcptest.io
   :target: https://mcptest.io/servers/exa/

Latest validation evidence

Validation status: Online when last tested
Validation checked at: Oct 26, 2026 at 11:06 AM
Validation detail: Negotiated stateful MCP 2025-11-25 at https://mcp.exa.ai/mcp; detected transport streamable-http; authentication none

Tool inventory

What the catalog validator observed on Oct 26, 2026 at https://mcp.exa.ai/mcp without credentials. The live server may have changed since; connect in the Playground for the current list.

Tools provided by Exa 2

Some details were sanitized for public display.

web_fetch_exaNo outputSchema
Read a webpage's full content as clean markdown. Use after web_search_exa when highlights are insufficient or to read any URL. Best for: Extracting full content from known URLs. Batch multiple URLs in one call. Returns: Clean text content and metadata from the page(s).
Input maxCharacters?: number, urls: array
- maxCharacters?: numberoptional
  Maximum characters to extract per page (default: 3000)
- urls: arrayrequired
  URLs to read. Batch multiple URLs in one call.
web_search_exaNo outputSchema
Search the web for any topic and get clean, ready-to-use content. Best for: Finding current information, news, facts, people, companies, or answering questions about any topic. Returns: Clean text content from top search results. Query tips: describe the ideal page, not keywords. "blog post comparing React and Vue performance" not "React vs Vue". Use category:people / category:company to search through Linkedin profiles / companies respectively. If highlights are insufficient, follow up with web_fetch_exa on the best URLs.
Input numResults?: number, objective: string, query: string
- numResults?: numberoptional
  Number of search results to return (default: 10).
- objective: stringrequired
  Goal for this search turn; say which documents should rank first, which should be excluded, and what specific facts or figures to pull from them.
- query: stringrequired
  Natural language search query. Should be a semantically rich description of the ideal page, not just keywords. Optionally include category:<type> (company, people) to focus results — e.g. 'category:people John Doe software engineer'.

Resources 1

tools_listapplication/json
List of available Exa tools and their descriptions

Prompts 1

web_search_help
Get help with web search using Exa
Arguments No parameters