Exa MCP Server
Exa's MCP server gives AI agents access to Exa's neural search engine built for AI: run web searches that return clean, ready-to-use content, fetch the full content of a specific page, run advanced filtered searches, and launch Exa research agents for multi-step, structured web research.
Install Exa
- Published by
- Official (vendor)
- Transport
- Remote (HTTP)
- Authentication
- API key
- Package
- exa-mcp-server
Claude Desktop, Claude Code and Cursor
Add this to the mcpServers object in your client's
config file, then restart the client.
{
"mcpServers": {
"exa": {
"type": "http",
"url": "https://mcp.exa.ai/mcp",
"headers": {
"x-api-key": "${EXA_API_KEY}"
}
}
}
}
VS Code uses a different key — show that config
Identical entry, filed under servers rather than
mcpServers. Put it in
.vscode/mcp.json for one workspace.
{
"servers": {
"exa": {
"type": "http",
"url": "https://mcp.exa.ai/mcp",
"headers": {
"x-api-key": "${EXA_API_KEY}"
}
}
}
}
Before it will answer
Unusually, this one works with no credential at all on the free tier — connect the URL and it answers. The key only lifts the rate limit, and goes in an x-api-key header rather than the usual bearer token. A local stdio alternative runs npx -y exa-mcp-server with EXA_API_KEY set.
Tools
agent_run
Run an Exa research agent for multi-step research and enrichment.
web_fetch_exa
Get the full content of a specific web page from a known URL.
web_search_advanced_exa
Advanced web search with full control over domains and dates.
web_search_exa
Search the web for any topic and get clean, ready-to-use content.