# Firecrawl

> Scrape, crawl, extract structured data, and search the web from an AI agent.

[Canonical HTML page](https://top-mcps.com/mcp/firecrawl) · [server.json](https://top-mcps.com/mcp/firecrawl.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": {
    "firecrawl": {
      "command": "npx",
      "args": [
        "-y",
        "firecrawl-mcp"
      ],
      "env": {
        "FIRECRAWL_API_KEY": "${FIRECRAWL_API_KEY}",
        "FIRECRAWL_API_URL": "${FIRECRAWL_API_URL}"
      }
    }
  }
}
```

### Claude Code — `CLI or .mcp.json`

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

```shell
# export FIRECRAWL_API_KEY=fc-your_key
# export FIRECRAWL_API_URL=https://api.firecrawl.dev
claude mcp add firecrawl -- npx -y firecrawl-mcp
```

### Cursor — `.cursor/mcp.json`

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

```json
{
  "mcpServers": {
    "firecrawl": {
      "command": "npx",
      "args": [
        "-y",
        "firecrawl-mcp"
      ],
      "env": {
        "FIRECRAWL_API_KEY": "${FIRECRAWL_API_KEY}",
        "FIRECRAWL_API_URL": "${FIRECRAWL_API_URL}"
      }
    }
  }
}
```

### VS Code — `.vscode/mcp.json`

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

```jsonc
{
  "servers": {
    "firecrawl": {
      "command": "npx",
      "args": [
        "-y",
        "firecrawl-mcp"
      ],
      "env": {
        "FIRECRAWL_API_KEY": "${FIRECRAWL_API_KEY}",
        "FIRECRAWL_API_URL": "${FIRECRAWL_API_URL}"
      }
    }
  }
}
```

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

Open via Cascade → hammer icon → Configure.

```json
{
  "mcpServers": {
    "firecrawl": {
      "command": "npx",
      "args": [
        "-y",
        "firecrawl-mcp"
      ],
      "env": {
        "FIRECRAWL_API_KEY": "${FIRECRAWL_API_KEY}",
        "FIRECRAWL_API_URL": "${FIRECRAWL_API_URL}"
      }
    }
  }
}
```

### Cline — `cline_mcp_settings.json`

Open via the Cline sidebar → MCP Servers → Edit.

```json
{
  "mcpServers": {
    "firecrawl": {
      "command": "npx",
      "args": [
        "-y",
        "firecrawl-mcp"
      ],
      "env": {
        "FIRECRAWL_API_KEY": "${FIRECRAWL_API_KEY}",
        "FIRECRAWL_API_URL": "${FIRECRAWL_API_URL}"
      }
    }
  }
}
```

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

Continue uses modelContextProtocolServers with a transport block.

```json
{
  "experimental": {
    "modelContextProtocolServers": [
      {
        "transport": {
          "type": "stdio",
          "command": "npx",
          "args": [
            "-y",
            "firecrawl-mcp"
          ],
          "env": {
            "FIRECRAWL_API_KEY": "${FIRECRAWL_API_KEY}",
            "FIRECRAWL_API_URL": "${FIRECRAWL_API_URL}"
          }
        }
      }
    ]
  }
}
```

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

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

```shell
# ~/.codex/config.toml
[mcp_servers.firecrawl]
command = "npx"
args = [
  "-y",
  "firecrawl-mcp",
]
env = { FIRECRAWL_API_KEY = "${FIRECRAWL_API_KEY}", FIRECRAWL_API_URL = "${FIRECRAWL_API_URL}" }
```

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

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

```jsonc
{
  "context_servers": {
    "firecrawl": {
      "command": {
        "path": "npx",
        "args": [
          "-y",
          "firecrawl-mcp"
        ]
      },
      "env": {
        "FIRECRAWL_API_KEY": "${FIRECRAWL_API_KEY}",
        "FIRECRAWL_API_URL": "${FIRECRAWL_API_URL}"
      }
    }
  }
}
```

### ChatGPT — `ChatGPT → Apps directory`

Firecrawl doesn't ship a hosted HTTPS endpoint today. ChatGPT supports remote MCP servers only — to use this server in ChatGPT you'll need to deploy it to a public HTTPS URL first (e.g. via Cloudflare Workers or Vercel) or wait for an official remote build.

```none

```

## At a glance

- **Maintainer:** Firecrawl
- **Transport:** stdio, Streamable HTTP, SSE
- **Auth model:** API key
- **Required secrets:** FIRECRAWL_API_KEY, FIRECRAWL_API_URL
- **Supported clients:** Claude, Cursor, Any MCP-compatible client, Firecrawl accounts
- **License:** MIT
- **Language:** JavaScript
- **Latest version:** latest
- **Last verified:** 2026-05-27
- **GitHub stars:** 6,393 (fetched 2026-06-02T11:55:52.255Z)
- **Score:** 90/100 (rubric 2026-04 — see https://top-mcps.com/about/methodology)
- **Source:** https://github.com/firecrawl/firecrawl-mcp-server

## Quick answer

**What it does.** Exposes Firecrawl operations as MCP tools: single-page scrape, depth-limited crawl, schema-driven extract, and web search with content fetching.

**Best for:**
- JS-rendered page scraping
- Multi-page crawls (docs sites)
- Schema-driven structured extraction
- Web search with content body
- Replacing custom Puppeteer infra

**Not for:**
- Sites with an existing API
- Volumes that exceed Firecrawl's rate limits
- Workflows that need raw HTML rather than cleaned markdown

## Description

The official Firecrawl MCP is the most-cited web-scraping MCP. Four tools: scrape (single page), crawl (multi-page with depth), extract (LLM-based structured extraction), and search (web search with optional content scraping).

## Why it matters

Firecrawl handles JS rendering, anti-bot evasion, and clean markdown output that most agent workflows want — without the host having to run Puppeteer themselves.

## Key features

- scrape + crawl + extract + search tools
- JS-rendered page support
- Clean markdown output
- stdio, HTTP, and SSE transports
- MIT, vendor-maintained

## FAQ

### Does it handle JavaScript?

Yes — Firecrawl renders JS server-side and returns the post-rendered DOM.

### How do I extract structured data?

The `extract` tool takes a schema (Zod-style or JSON Schema) and an LLM does the extraction. Useful for pulling fields out of listings.

### Self-hosted option?

Yes — set `FIRECRAWL_API_URL` to your own Firecrawl deployment.

## Changelog

- **2026-05-27** — Refreshed install snippets and fact sheet; verified for 2026.
- **2025-02-15** — Initial directory listing.