Skip to content

MCP server

Firecrawl MCP server

Firecrawl's official MCP server lets AI scrape, map, crawl and search the web and get clean, AI-ready content back.

At a glance

Maintained by
Firecrawl
Tools exposed
10 listed below
Reviewed
September 24, 2026

Why it matters

What it does and why it matters

Firecrawl turns websites into clean content that AI can use. Its official MCP server lets an assistant scrape a page, map a site, crawl it, search the web, parse documents and run an agent that interacts with pages. Results come back as clean markdown or structured data rather than raw HTML.

That makes it a practical tool for building knowledge bases from your own sites, monitoring public pages and gathering research material.

The hosted server now uses a newer version of the endpoint and offers a limited free tier without a key, with a separate sign-in option. Usage beyond that draws on account credits, and the interact tool can take actions on pages. We keep crawls to known sites, set page limits and check credit usage after each job.

Use cases

Best business use cases

Knowledge base from your site

Turn your help center into content for an assistant. See knowledge base search.

Site maps

Map a site before a redesign or migration.

Price and page monitoring

Scrape public pages on a schedule. See price monitoring.

Research

Search and scrape sources for a brief.

Document parsing

Parse documents into clean text.

Tools

Available tools

Tool names as published by the maintainer at review time. Versions change, so check the repository for the current list.

Read and write: Mostly read: fetches and extracts web content; interact can take page actions; uses account credits.

  • firecrawl_scrape

    Get a known URL's content as markdown or schema-shaped JSON.

  • firecrawl_map

    Discover URLs on a site without fetching content.

  • firecrawl_crawl

    Fetch content from many pages under a site.

  • firecrawl_check_crawl_status

    Check the status of a crawl job.

  • firecrawl_search

    Search the web and optionally scrape the results.

  • firecrawl_parse

    Parse an uploaded file into usable content.

  • firecrawl_agent

    Run an autonomous research agent that returns structured data.

  • firecrawl_agent_status

    Poll an agent job for its results.

  • firecrawl_interact

    Click, type or move around a page before reading it.

  • firecrawl_credit_usage

    Check remaining credits and usage.

Workflows

Example workflows

Load a help center into an assistant

  1. 1

    Map your help center.

  2. 2

    Crawl it with a page limit.

  3. 3

    Review the clean markdown output.

  4. 4

    Load it into your knowledge base.

  5. 5

    Test the assistant with real questions.

Check competitor page changes

  1. 1

    List the pages to watch.

  2. 2

    Scrape them each week.

  3. 3

    Compare with last week's content.

  4. 4

    Summarize what changed.

  5. 5

    Send the summary to the team.

Prompts

Example prompts

  • “Map our help center and list all article pages.”

  • “Scrape this page and return the product specifications as JSON.”

  • “How many credits have we used this month?”

Setup

Installation and configuration

Examples for Claude Desktop and Claude Code. Replace the placeholder values with your own keys, and store secrets outside version control.

Claude Desktop: claude_desktop_config.json

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

Claude Code: terminal

claude mcp add --transport http firecrawl https://mcp.firecrawl.dev/v2/mcp-oauth
  • Hosted server: streamable-http; https://mcp.firecrawl.dev/v2/mcp; https://mcp.firecrawl.dev/v2/mcp-oauth; Keyless (limited), OAuth, or header Authorization: Bearer <FIRECRAWL_API_KEY>.
  • Local server: stdio; firecrawl-mcp.
  • Needs: FIRECRAWL_API_KEY (local; optional on hosted keyless tier).

Security

Security considerations

  • Site terms

    Crawl only sites you are allowed to use.

  • Credits

    Watch credit usage and set limits.

  • Untrusted content

    Treat scraped text as data, not instructions.

  • Interact tool

    Review any actions the interact tool takes on pages.

Related servers

Servers that pair well with it

Apify

Apify's official MCP server lets AI find and run ready-made scrapers and automation tools, called Actors, and read their results.

Tavily

Tavily's official MCP server gives AI web search, content extraction, site mapping and crawling, all built for AI use.

Playwright

Microsoft's Playwright MCP server lets AI drive a real browser: open pages, read them, click, type, fill forms and take screenshots.

Keep exploring

FAQ

Questions people ask us

Have a question that is not here? Ask us directly.

Start a project

Want this connected to your own systems? We build AI workflows with MCP.

Send a short brief. We reply with questions, a suggested plan and an estimate you can compare with other offers.

Your privacy choices

We use necessary storage to run this site. With your permission we also use Google Analytics to see which pages help people, and load maps from Google. You can change this at any time. Read the cookie policy.