Skip to content
Firecrawl logo

Firecrawl

Verified

Brings Firecrawl web scraping and search to MCP AI agents.

MCP ServerSearch & WebLocal (stdio) 6.6k
View on GitHub
Updated 2026-06-15

What is the Firecrawl MCP server?

This official MCP server exposes Firecrawl functionality for clean, agent-ready web context. It handles search, scraping, page interactions, and deep research while managing rate limits and retries automatically.

Users can run it via npx with an API key or configure it in editors like Cursor, VS Code, and Windsurf. SSE and Streamable HTTP modes are also supported for flexible deployments.

Install & connect

Add this to your MCP client config. Pick your client below and copy.

{
  "mcpServers": {
    "firecrawl-mcp-server": {
      "command": "npx",
      "args": [
        "-y",
        "firecrawl-mcp"
      ]
    }
  }
}

Package: firecrawl-mcp (npm)

Other ways to install

npx

npx
npx -y firecrawl-mcp

Example prompts

Once connected, try asking your AI client:

Search the web for the latest AI model releases and summarize key features
Scrape https://example.com/docs and extract all API endpoint details
Research current trends in renewable energy using deep web search
Navigate to a product page and extract pricing and availability information

Security & permissions

Requires a Firecrawl API key passed via the FIRECRAWL_API_KEY environment variable. Runs locally via stdio or HTTP and sends requests to Firecrawl's cloud or self-hosted endpoints.

What you can do with Firecrawl

Web Research

Perform autonomous deep research by searching and scraping multiple sources for comprehensive answers.

Data Extraction

Scrape any URL to obtain clean structured data suitable for analysis or further processing.

Page Automation

Interact with live web pages by clicking, navigating, and performing actions within the agent workflow.

How to use Firecrawl

  1. 1Obtain a Firecrawl API key from firecrawl.dev
  2. 2Install via npx with the API key in the environment
  3. 3Add the server configuration to your client (Cursor, VS Code, etc.)
  4. 4Restart the client and refresh the MCP server list
  5. 5Use prompts that involve web search or scraping to invoke the tools

Firecrawl: pros & cons

Pros

  • +Official integration with robust retry and rate-limiting logic
  • +Supports multiple transports including stdio and Streamable HTTP
  • +Works across popular editors like Cursor and VS Code
  • +Enables autonomous deep research beyond simple scraping

Cons

  • Requires a paid or free-tier Firecrawl API key
  • Tool names and parameters are not documented in the README
  • Depends on external Firecrawl service availability
Did you find this helpful?

Frequently asked questions

Yes, the FIRECRAWL_API_KEY environment variable must be set to run the server.

User reviews

Verified reviews from the community shape this listing's rating.

Loading reviews…

Sign in to review

Promote Firecrawl

Add this badge to your website, or share the tool.

DFeatured on DhanasviFirecrawl 1