# Exa

> Neural web search and page fetch for AI agents — official remote MCP at mcp.exa.ai/mcp.

[Canonical HTML page](https://top-mcps.com/mcp/exa) · [server.json](https://top-mcps.com/mcp/exa.json) · [methodology](https://top-mcps.com/about/methodology)

## Install

### Claude Desktop — `claude_desktop_config.json`

Paste under mcpServers. Fully quit and reopen Claude after editing.

```json
{
  "mcpServers": {
    "exa": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://mcp.exa.ai/mcp"
      ],
      "env": {
        "EXA_API_KEY": "${EXA_API_KEY}"
      }
    }
  }
}
```

### Claude Code — `CLI or .mcp.json`

Run from your repo. Commit .mcp.json to share with your team.

```shell
# export EXA_API_KEY=your_exa_api_key
claude mcp add exa -- npx -y mcp-remote https://mcp.exa.ai/mcp
```

### Cursor — `.cursor/mcp.json`

Global path: ~/.cursor/mcp.json. Reload window after editing.

```json
{
  "mcpServers": {
    "exa": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://mcp.exa.ai/mcp"
      ],
      "env": {
        "EXA_API_KEY": "${EXA_API_KEY}"
      }
    }
  }
}
```

### VS Code — `.vscode/mcp.json`

VS Code uses the "servers" key (not "mcpServers").

```jsonc
{
  "servers": {
    "exa": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://mcp.exa.ai/mcp"
      ],
      "env": {
        "EXA_API_KEY": "${EXA_API_KEY}"
      }
    }
  }
}
```

### Windsurf — `~/.codeium/windsurf/mcp_config.json`

Open via Cascade → hammer icon → Configure.

```json
{
  "mcpServers": {
    "exa": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://mcp.exa.ai/mcp"
      ],
      "env": {
        "EXA_API_KEY": "${EXA_API_KEY}"
      }
    }
  }
}
```

### Cline — `cline_mcp_settings.json`

Open via the Cline sidebar → MCP Servers → Edit.

```json
{
  "mcpServers": {
    "exa": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://mcp.exa.ai/mcp"
      ],
      "env": {
        "EXA_API_KEY": "${EXA_API_KEY}"
      }
    }
  }
}
```

### Continue — `~/.continue/config.json`

Continue uses modelContextProtocolServers with a transport block.

```json
{
  "experimental": {
    "modelContextProtocolServers": [
      {
        "transport": {
          "type": "stdio",
          "command": "npx",
          "args": [
            "-y",
            "mcp-remote",
            "https://mcp.exa.ai/mcp"
          ],
          "env": {
            "EXA_API_KEY": "${EXA_API_KEY}"
          }
        }
      }
    ]
  }
}
```

### Codex CLI — `~/.codex/config.toml`

Codex uses TOML. Each server is a [mcp_servers.<name>] subtable.

```shell
# ~/.codex/config.toml
[mcp_servers.exa]
command = "npx"
args = [
  "-y",
  "mcp-remote",
  "https://mcp.exa.ai/mcp",
]
env = { EXA_API_KEY = "${EXA_API_KEY}" }
```

### Zed — `~/.config/zed/settings.json`

Zed calls them "context_servers". Settings live-reload on save.

```jsonc
{
  "context_servers": {
    "exa": {
      "command": {
        "path": "npx",
        "args": [
          "-y",
          "mcp-remote",
          "https://mcp.exa.ai/mcp"
        ]
      },
      "env": {
        "EXA_API_KEY": "${EXA_API_KEY}"
      }
    }
  }
}
```

### ChatGPT — `ChatGPT → Settings → Connectors (remote URL)`

Paste this into Settings → Connectors → Add custom connector. Requires ChatGPT Pro / Team / Enterprise / Edu with Developer mode enabled. ChatGPT supports remote HTTPS MCP servers only — stdio servers must be hosted on a public HTTPS endpoint first.

```json
{
  "name": "Exa",
  "transport": "http",
  "url": "https://mcp.exa.ai/mcp"
}
```

## At a glance

- **Maintainer:** Exa
- **Transport:** Streamable HTTP, stdio
- **Auth model:** API key
- **Required secrets:** EXA_API_KEY
- **Supported clients:** Claude, Cursor, Any MCP-compatible client, Exa accounts
- **License:** MIT
- **Language:** TypeScript
- **Latest version:** latest
- **Last verified:** 2026-05-27
- **GitHub stars:** 4,487 (fetched 2026-06-02T11:55:52.245Z)
- **Score:** 89/100 (rubric 2026-04 — see https://top-mcps.com/about/methodology)
- **Source:** https://github.com/exa-labs/exa-mcp-server

## Quick answer

**What it does.** Exposes Exa neural search and content retrieval as MCP tools: web_search_exa, web_fetch_exa, and an optional advanced search tool.

**Best for:**
- LLM-optimised web search
- Fetching a URL's contents for context
- Research and fact-checking
- Finding docs by description
- Zero-config remote install

**Not for:**
- Real-time news (try Brave Search)
- Free-tier-only workflows (paid plans for sustained use)
- Needs outside Exa's API surface

## Description

The official Exa MCP exposes Exa's neural web search index plus a page-fetch tool over a hosted remote endpoint. API-key auth, no install, designed for AI agent retrieval workflows.

## Why it matters

Exa was purpose-built for AI retrieval — its index ranks by relevance to natural-language queries rather than keyword match. The remote MCP is the cleanest way to plug that into Claude or Cursor.

## Key features

- Hosted at mcp.exa.ai/mcp — no install
- Neural search ranked for LLM use
- web_fetch_exa for URL → content
- API-key auth via env var
- MIT, vendor-maintained

## FAQ

### How is it different from Brave Search?

Exa is neural and ranks for LLM relevance; Brave is keyword-driven with strong real-time freshness. Pair them if you need both.

### Free tier?

Yes — there is a free tier on Exa's dashboard. Sustained use needs a paid plan.

### Local install option?

Yes — the npm `mcp-remote` wrapper proxies the remote MCP via stdio for clients that prefer local transports.

## Changelog

- **2026-05-27** — Refreshed install snippets and fact sheet; verified for 2026.
- **2025-03-10** — Initial directory listing.