Firecrawl MCP review

Q: How do I install Firecrawl MCP?

Run env FIRECRAWL_API_KEY=fc-YOUR_API_KEY npx -y firecrawl-mcp. It needs Node.js >=22.0.0. The install section above has the other options from the project's README.

Firecrawl MCP is the official Model Context Protocol server that lets MCP clients search, scrape, crawl, map, parse, and interact with the web through Firecrawl.

Updated Nov 1, 2026. Review written Oct 22, 2026 from the project's own sources; usage data refreshed every Monday.

TL;DR

Firecrawl MCP brings the Firecrawl API to MCP-compatible clients such as Cursor, Windsurf, and VS Code. It returns web content as clean Markdown or structured data. It is for developers building AI agents that need live web data. It offers a hosted endpoint with a keyless tier, local npx and Docker options, and support for self-hosted Firecrawl.

7.7kGitHub stars
+55new stars, week of Oct 25
~179.6kdownloads a week, −5% vs 4-week avg
#51in MCP and agent tools this week

Firecrawl MCP at a glance

Category: MCP and agent tools
Maintained by: Firecrawl (organization, @firecrawl on GitHub)
License: MIT
Language: JavaScript, TypeScript, Dockerfile
Repository: github.com/firecrawl/firecrawl-mcp-server
Website: firecrawl.dev
npm package: firecrawl-mcp 3.33.4
Latest release: firecrawl-mcp 3.33.4 on npm, Oct 20, 2026
First published: Dec 6, 2024

What Firecrawl MCP does

Firecrawl MCP scrapes any URL into clean Markdown or into structured JSON that matches a schema you supply.
Firecrawl MCP searches the web, plus indexes for developer sources, US government sources, and scientific papers such as PubMed and arXiv.
Firecrawl MCP crawls a site for content from many pages, or maps a site to discover its URLs without fetching content.
Firecrawl MCP interacts with pages in a real browser: it can click, type, fill forms, and navigate dynamic pages.
Firecrawl MCP parses PDFs and other documents.
Firecrawl MCP runs Agent jobs for multi-source research and monitors pages for content changes.
Firecrawl MCP reports current and historical Firecrawl credit usage, by month or day, optionally per endpoint or API key.

How it works

Firecrawl MCP runs as a hosted endpoint at https://mcp.firecrawl.dev/v2/mcp, or locally through npx or Docker. Locally it uses stdio by default, with optional Streamable HTTP at localhost:3000/mcp. The server forwards requests to the Firecrawl cloud API using an API key or OAuth token. You can also point it at a self-hosted Firecrawl API with FIRECRAWL_API_URL.

How to install and start using Firecrawl MCP

With npx

env FIRECRAWL_API_KEY=fc-YOUR_API_KEY npx -y firecrawl-mcp

Global npm install

npm install -g firecrawl-mcp

With Docker

docker run -i --rm -e FIRECRAWL_API_KEY=fc-YOUR_API_KEY ghcr.io/firecrawl/firecrawl-mcp-server:latest

Requirements

Node.js 22.0.0 or later to run the npm package.
A Firecrawl API key is required for the cloud API. It is optional for a self-hosted instance.
Cursor users need Cursor version 0.45.6 or later.

First steps

Create a Firecrawl account and get an API key from https://www.firecrawl.dev/app/api-keys.
Add the server to your MCP client config, using npx with the args -y firecrawl-mcp and FIRECRAWL_API_KEY in env.
Refresh your client's MCP server list to see the Firecrawl tools.
Ask your agent for web content. For example, use firecrawl_scrape for a known URL or firecrawl_search for a query.

Who Firecrawl MCP is for

Developers giving coding agents in Cursor, Windsurf, or VS Code access to live web content.
Agent builders who need web pages as clean Markdown or schema-matched JSON.
Researchers searching scientific papers, government sources, or developer sources from an MCP client.
Teams running a self-hosted Firecrawl instance who want MCP access to it.

Limitations to know

On the keyless hosted tier, only scrape, search, and parse work, with rate limits. Crawl, map, and agent need a key.
Each firecrawl_interact call runs one turn to completion. You cannot drive a browser session step by step within a single call.
The full profile registers many tools, which can strain clients with tool-slot limits. Smaller endpoints exist for those clients.
Using scrapeOptions with search fetches every result page and bills each one as a scrape.

Is Firecrawl MCP growing? Usage data

In the week of Oct 25–31, 2026, Firecrawl MCP ranked #330 on the Devtoolwatch Board and #51 in MCP and agent tools, with a momentum score of 26 out of 100. Its main package had about 179.6k downloads that week, −5% against its own 4-week average.

New GitHub stars per week for the last 12 weeks, from GitHub's stargazer history (today's stargazers by the week they starred).

Show the weekly numbers

New GitHub stars per week for Firecrawl MCP
Week of	New stars
Aug 9, 2026	65
Aug 16, 2026	54
Aug 23, 2026	55
Aug 30, 2026	57
Sep 6, 2026	41
Sep 13, 2026	47
Sep 20, 2026	30
Sep 27, 2026	36
Oct 4, 2026	43
Oct 11, 2026	58
Oct 18, 2026	57
Oct 25, 2026	55

On Hacker News

How to Build an AI Software Factory: Agents That Open, Review, and Merge PRs, 31 points and 14 comments (Sep 11, 2026)
Show HN: Llms.txt Generator – Turn websites into a text file to feed to any LLM, 28 points and 6 comments (Nov 21, 2024)
Show HN: API to turn entire websites into Markdown, 23 points and 4 comments (Apr 16, 2024)
Turn entire websites into LLM-ready data, 16 points and 11 comments (Apr 13, 2024)
Show HN: From SERP to LLM-Ready Data Markdown in Seconds with Firesearch, 14 points and 0 comments (Apr 25, 2024)

The Board ranks tools by four signals each week: new stars, new stars for the repo's size, added downloads and download growth beyond the market. How the ranking works.

Firecrawl MCP alternatives

Other MCP servers and agent tools on the Board in the week of Oct 25–31, 2026, by momentum:

MCP and agent tools ranked by momentum, the week of Oct 25–31, 2026
Rank	Tool	Stars	Downloads/wk
#1	REAREA is an MCP server and CLI that lets coding agents reverse engineer native binaries, JavaScript and Electron apps, .NET assemblies,…	415.5k	760.6k
#2	ha-mcpha-mcp is an unofficial Model Context Protocol server that lets AI assistants control and manage Home Assistant using natural language.	5.4k	118.1k
#3	TregTreg is a registry and proxy that gives AI agents one token for a catalog of paid third-party API endpoints and team-shared tools.	6.9k	6.7k
#4	Chrome DevTools MCPChrome DevTools MCP is an MCP server and CLI that lets coding agents control and inspect a live Chrome browser using Chrome DevTools.	54.6k	2.7M
#5	codebase-memory-mcpHigh-performance code intelligence MCP server. Indexes codebases into a persistent knowledge graph — average repo in milliseconds. 162…	47.9k	–
#6	open-computer-use👾 Open Computer Use – Open-Source Alternative to Codex Computer Use	2.8k	–

All MCP servers and agent tools, ranked

Recent Firecrawl MCP releases

v3.2.1, Sep 26, 2025
v3.2.0, Sep 16, 2025
v3.0.0, Sep 11, 2025
v2.0.0, Aug 23, 2025
v1.12.0, Jul 3, 2025

Firecrawl MCP FAQ

Is Firecrawl MCP free and open source?

Yes. Firecrawl MCP's source code is on GitHub (firecrawl/firecrawl-mcp-server) under the MIT license, so you can use it for free. Its website also links to a pricing page (https://firecrawl.dev/pricing) for any paid plans.

How do I install Firecrawl MCP?

Run env FIRECRAWL_API_KEY=fc-YOUR_API_KEY npx -y firecrawl-mcp. It needs Node.js >=22.0.0. The install section above has the other options from the project's README.

Is Firecrawl MCP still maintained?

Yes. The firecrawl/firecrawl-mcp-server repository was last updated on Oct 21, 2026, and the latest release, firecrawl-mcp 3.33.4 on npm, came out on Oct 20, 2026.

How popular is Firecrawl MCP?

Firecrawl MCP has 7.7k GitHub stars as of Nov 1, 2026. It gained 55 new stars in the week of Oct 25–31, 2026, against a 12-week average of 50 a week. Its main package had about 179.6k downloads that week, −5% against its own 4-week average. That made it #51 among MCP servers and agent tools on the Devtoolwatch Board.

What are the best alternatives to Firecrawl MCP?

Other MCP servers and agent tools on the Devtoolwatch Board in the week of Oct 25–31, 2026, by momentum: REA (#1), ha-mcp (#2), Treg (#3), Chrome DevTools MCP (#4), codebase-memory-mcp (#5). The alternatives table above has their numbers.

What is Firecrawl MCP?

Firecrawl MCP is the official Model Context Protocol server for Firecrawl. It lets MCP-compatible clients search, scrape, crawl, map, parse, and interact with the live web. It returns clean Markdown or structured data.

What is Firecrawl MCP used for?

Firecrawl MCP is used to give AI agents web data. Uses include scraping known URLs, searching from a query, crawling or mapping sites, filling forms in a real browser, running multi-source research jobs, and monitoring pages for changes.

How do I set up Firecrawl MCP in VS Code?

Firecrawl MCP can be installed in VS Code with the one-click install buttons in the README. You can also add a JSON block to your User Settings (JSON) or to .vscode/mcp.json. The block runs npx -y firecrawl-mcp and prompts for your Firecrawl API key.

Can Firecrawl MCP be used without an API key?

Firecrawl MCP offers a hosted keyless endpoint at https://mcp.firecrawl.dev/v2/mcp. On that tier, scrape, search, and parse work with rate limits. Other tools such as crawl, map, and agent still need a key or OAuth.

Does Firecrawl MCP work with self-hosted Firecrawl?

Firecrawl MCP works with self-hosted Firecrawl. Set FIRECRAWL_API_URL to your instance. In that case the API key is optional.

Should I use Firecrawl MCP for long interactive browser sessions?

Firecrawl MCP is not designed for long interactive browser sessions. Each firecrawl_interact call runs one prompt or code turn and then returns. Sessions persist via scrapeId and end with firecrawl_interact_stop. The README suggests considering something else if you need your own step-by-step control.

Sources

Firecrawl MCP README badge

Maintainers can show the tool's weekly Board rank in their README:

[![Devtoolwatch](https://index.devtoolwatch.com/badge/firecrawl-mcp.svg)](https://index.devtoolwatch.com/tools/firecrawl-mcp/)

How this page is made: Devtoolwatch writes each review from the project's README, releases, package metadata and website, with an AI model, then checks every claim against the source it cites and drops what the sources don't support. The numbers come from Devtoolwatch's daily GitHub snapshots, GitHub's stargazer history, and the npm and PyPI download APIs. Spotted an error? Email [email protected].