Firecrawl Local

You can run a local MCP server that talks to your self-hosted Firecrawl instance to perform web scraping, crawling, and site mapping. This setup lets you operate Firecrawl-related tasks from your own environment without exposing external APIs while keeping everything self-contained.

How to use

Connect to the local MCP server from your MCP client (Claude Desktop, Cline, or another MCP client) and use the available tools to scrape pages, crawl sites, or map URLs. The server is designed to work with a Firecrawl instance running on your network, using a configurable Firecrawl URL.

How to install

Prerequisites: you need Node.js 18 or newer installed on your system.

Install dependencies and build the MCP server package.

npm install
npm run build

Configuration

The MCP server connects to your Firecrawl instance using an environment variable named FIRECRAWL_URL. By default, it connects to http://localhost:3002.

Set the Firecrawl URL in your MCP client configuration to point to your local Firecrawl instance. You can use either the absolute path example for a direct node invocation or the simpler relative path with a working directory.

{
  "mcpServers": {
    "firecrawl_local": {
      "command": "node",
      "args": ["dist/index.js"],
      "env": {
        "FIRECRAWL_URL": "http://localhost:3002"
      }
    }
  }
}

Testing

Test the server functionality to ensure the tool list and a sample scrape operation work as expected.

node test.js

Example usage with a configured client

After configuration, you can issue natural-language like commands through your MCP client to perform actions such as scraping a page, crawling a site, or mapping URLs.

Available tools

firecrawl_scrape

Scrape a single webpage and return its content in markdown format. Parameters: url (required), formats (default: ["markdown"]), onlyMainContent (default: true), includeTags, excludeTags.

firecrawl_crawl

Crawl a website starting from a URL and return content from multiple pages. Parameters: url (required), includes, excludes, maxDepth (default: 2), limit (default: 10), allowBackwardLinks (default: false), allowExternalLinks (default: false).

firecrawl_crawl_status

Check the status of a crawl job. Parameters: jobId (required).

firecrawl_map

Map a website to get a list of all accessible URLs. Parameters: url (required), search, ignoreSitemap (default: false), includeSubdomains (default: false), limit (default: 5000).