Firecrawl

Scrape pages, crawl sites and search, returned as clean Markdown.

Started in our test
★ 7.7K GitHub stars
128.6K downloads a week
Last commit 26 Oct 2026

At a glance

Firecrawl's server sends requests to the Firecrawl API, which renders pages, strips navigation and ads, and returns Markdown or structured JSON. It can also map a site's URLs, crawl them, and run a web search and scrape the results in one go.

Because the work happens on Firecrawl’s side, it copes with JavaScript-heavy pages without a browser on your machine.

Made by: Firecrawl (Mendable)
Licence: MIT
Latest version: 3.34.1 (22 Oct 2026)
npm package: firecrawl-mcp
GitHub stars: 7.7K
Downloads: 128.6K a week
Last commit: 26 Oct 2026
Tools: 28
Runs on: Your computer

Our check: We ran this exact config on 26 Oct 2026: it started and answered with 28 tools. firecrawl-mcp on npm is published from firecrawl/firecrawl-mcp-server.

What's in the pod: its tools

We found 28 tools when we started it on 26 Oct 2026. 17 only read. The rest can change things, so keep your app's "ask before each tool call" setting on for them.

Runs commands1 tool · Can start programs or run code.

Tool	What it is for
`firecrawl_monitor_run`	Runs a monitor's check right away.

Changes things10 tools · Can create, edit, send or delete.

Tool	What it is for
`firecrawl_agent`	Hands a research goal to Firecrawl's own agent, which browses on its own and returns structured findings.
`firecrawl_crawl`	Walks a whole site or section and fetches every page it finds, within limits you set.
`firecrawl_feedback`	Sends a rating on any Firecrawl result back to the service.
`firecrawl_interact`	Clicks, types or navigates on a live page, for content that only appears after an action.
`firecrawl_interact_stop`	Closes a browser session opened by firecrawl_interact.
`firecrawl_monitor_create`	Sets up a page to be re-checked on a schedule with change alerts.
`firecrawl_monitor_delete`	Removes a monitor.
`firecrawl_monitor_update`	Changes a monitor's schedule or target.
`firecrawl_scrape`	Fetches one URL and returns it as Markdown, HTML, a screenshot or JSON that fits a schema you give.
`firecrawl_search_feedback`	Rates a past search so Firecrawl can improve it, sometimes refunding a credit.

Only reads17 tools · Looks things up and changes nothing.

Tool	What it is for
`firecrawl_agent_status`	Checks whether an agent job has finished and collects its results.
`firecrawl_check_crawl_status`	Reports progress on a crawl and returns the pages gathered so far.
`firecrawl_credit_usage`	Shows how many credits you have left and what used them, by day or month.
`firecrawl_developer_search`	Searches public code, GitHub issues, pull requests and library docs for programming answers.
`firecrawl_find_tools`	Browses Firecrawl's catalogue of ready-made data providers.
`firecrawl_gov_search`	Searches US federal, state and local laws, regulations and court opinions.
`firecrawl_map`	Lists the URLs on a site without downloading their content.
`firecrawl_monitor_check`	Shows the result of one scheduled check, including what changed.
`firecrawl_monitor_checks`	Lists past checks for a monitor.
`firecrawl_monitor_get`	Shows one monitor's settings.
`firecrawl_monitor_list`	Lists your monitors.
`firecrawl_parse`	Turns an uploaded document, such as a PDF, into clean text.
`firecrawl_research_inspect_paper`	Gives a quick look at one paper's details and sections.
`firecrawl_research_read_paper`	Reads passages from a paper's full text.
`firecrawl_research_related_papers`	Follows citations to papers linked to one you already have.
`firecrawl_research_search_papers`	Searches biomedical and arXiv research papers.
`firecrawl_search`	Searches the web and returns ranked results with excerpts, optionally scraping each result too.

Grouped and described by us. The README on GitHub lists the inputs each tool takes.

Set it up

Pick your app for the exact snippet and where it goes.

Where it goes: claude_desktop_config.json (Settings → Developer → Edit Config)

Claude Desktop

{
  "mcpServers": {
    "firecrawl": {
      "command": "npx",
      "args": [
        "-y",
        "firecrawl-mcp"
      ],
      "env": {
        "FIRECRAWL_API_KEY": "fc-your-key"
      }
    }
  }
}

Where it goes: Your terminal, in the project folder

Claude Code

claude mcp add-json firecrawl '{"command":"npx","args":["-y","firecrawl-mcp"],"env":{"FIRECRAWL_API_KEY":"fc-your-key"}}'

Where it goes: ~/.cursor/mcp.json, or .cursor/mcp.json in a project

Cursor

{
  "mcpServers": {
    "firecrawl": {
      "command": "npx",
      "args": [
        "-y",
        "firecrawl-mcp"
      ],
      "env": {
        "FIRECRAWL_API_KEY": "fc-your-key"
      }
    }
  }
}

Where it goes: .vscode/mcp.json in your project

VS Code

{
  "servers": {
    "firecrawl": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "firecrawl-mcp"
      ],
      "env": {
        "FIRECRAWL_API_KEY": "fc-your-key"
      }
    }
  }
}

Where it goes: ~/.codeium/windsurf/mcp_config.json

Windsurf

{
  "mcpServers": {
    "firecrawl": {
      "command": "npx",
      "args": [
        "-y",
        "firecrawl-mcp"
      ],
      "env": {
        "FIRECRAWL_API_KEY": "fc-your-key"
      }
    }
  }
}

New to this? See the setup guides for Claude Code, Cursor, Claude Desktop and ChatGPT.

Settings

Names are exactly as the project spells them. Required ones are marked.

Setting	What it does
`FIRECRAWL_API_KEY`Environment variable · required	Your key from firecrawl.dev. Needed for the cloud service unless you use the hosted keyless tier or OAuth.`fc-xxxxxxxx`
`FIRECRAWL_API_URL`Environment variable	Address of your own self-hosted Firecrawl. The key becomes optional if your instance has no auth.`https://firecrawl.your-domain.com`
`https://mcp.firecrawl.dev/v2/mcp`Config	Hosted address. Scrape, search and parse work without a key on the free tier; send a Bearer key header for everything else.
`https://mcp.firecrawl.dev/v2/mcp-oauth`Config	Hosted address that signs in to your Firecrawl account through your client, no key to copy.
`FIRECRAWL_OAUTH_TOKEN`Environment variable	A fixed OAuth access token (starting fco_) for the local stdio mode instead of an API key.
`FIRECRAWL_NO_SEARCH_FEEDBACK`Environment variable	Set to 1 to hide the feedback tools and keep the tool list shorter.`1`

What it can touch

It can read

Public web pages, search results, papers, code and government records
Your credit balance

It can change

Nothing on your machine
Uses Firecrawl credits for each scrape, crawl, search or agent run
Creates and deletes page monitors in your account

Safer setup: Set a limit and include or exclude paths on every crawl so a single request cannot burn through credits.

Before you start

A Firecrawl account and API key (there is a free tier)
Node.js 18 or newer

If it does not work

Crawl, map or agent say a key is needed: The keyless tier only covers scrape, search and parse. Add an API key or connect through the OAuth address.
Rate limit exceeded: Wait and retry, or move to a paid plan. Check what is left with firecrawl_credit_usage.
A page comes back empty: It may need a click or scroll first. Use firecrawl_interact, or ask scrape to wait before capturing.
A crawl runs far longer than expected: Cap it with limit and maxDiscoveryDepth, or use firecrawl_map first to see how big the site is.

Peas in a pod: similar servers

How Firecrawl lines up against other browser and web data servers on the shelf.

Server	Tools	Runs on	Licence	Stars
Firecrawl	28	Your computer	MIT	7.7K
Exa	2	Your computer or hosted	–	5.2K
Bright Data	32	Your computer	MIT	2.7K
SearXNG	4	Your computer	MIT	1.3K
Playwright MCP	25	Your computer	Apache-2.0	38.7K