Firecrawl Web Scraping & Search

firecrawl

Scrape, search, map, crawl, and extract structured data from the live web via the official Firecrawl MCP server.

Installs: 6.6k
Namespace: firecrawl
Version: >=3.0.0 <4.0.0
Last attach: reviewed1 months ago

Add to your mcp.json

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

Open in VS CodeDon't have it? Install FrootAI for VS Code

Attach in a FrootAI play (fai-manifest)

mcp_scope:
  attached:
    - firecrawl

Authentication

api-keyoauth-token

create a Firecrawl account and set FIRECRAWL_API_KEY to your `fc-…` key (firecrawl.dev/app/api-keys); or point FIRECRAWL_API_URL at a self-hosted instance.

Required credential: FIRECRAWL_API_KEY

Environment variables

Variable	Required	Auth mode	Description
FIRECRAWL_API_KEY	yes	api-key	Firecrawl API key (`fc-…`) for the cloud API (the default). Optional only when pointing at a self-hosted instance via FIRECRAWL_API_URL. Pass via env, never inline in args (doctrine #6).
FIRECRAWL_API_URL	no	any	Custom API endpoint for a self-hosted Firecrawl instance (e.g. https://firecrawl.your-domain.com). Defaults to the Firecrawl cloud API.
FIRECRAWL_OAUTH_TOKEN	no	oauth-token	Static Firecrawl OAuth access token (`fco_…`) — an alternative to FIRECRAWL_API_KEY for stdio. Use access tokens only, never refresh tokens (`fcr_…`).

Sample tools

firecrawl.firecrawl_scrape
Scrape a single URL into clean structured data or markdown.
firecrawl.firecrawl_search
Search the web and optionally scrape result content.
firecrawl.firecrawl_map
Discover all indexed URLs on a website.
firecrawl.firecrawl_crawl
Start an async crawl of a site and extract content from its pages.
firecrawl.firecrawl_check_crawl_status
Check the status of a crawl job.
firecrawl.firecrawl_extract
Extract structured data from web pages using an LLM and a schema.
firecrawl.firecrawl_batch_scrape
Scrape multiple known URLs efficiently in one batch.

Used in recipes

FAI cookbook recipes that attach firecrawl.

Destructive-action audit

No destructive tools — audited read-only

Destructive-action audit (X2.18): no destructive tools — Firecrawl scrapes/searches/extracts public web content and does not mutate your resources.

Known limitations

Requires a Firecrawl API key (FIRECRAWL_API_KEY) for the default cloud API; a self-hosted instance via FIRECRAWL_API_URL can run without one.
Scrape, search, crawl, and extract consume Firecrawl credits — deep crawls and the autonomous agent draw down quota faster; credit-warning thresholds are configurable.
Crawl and agent jobs are asynchronous — they return a job ID and you poll firecrawl_check_crawl_status / firecrawl_agent_status for results, which can take minutes.

Spec contributors

Maintainers who landed commits onmcp-specs/firecrawl.json.

pspsbali

Freshness

When this spec was last reviewed by a maintainer and last machine-validated against the live server.

Spec last reviewed: 1 months ago(2026-06-25)
Auto-validated: never