The server is open source under an MIT license and free to run; calls to Firecrawl's cloud API consume credits, and a FIRECRAWL_API_KEY is required. A free tier with starter credits is available, and you can self-host Firecrawl to avoid cloud usage.

Does it support remote/OAuth?

Yes. Firecrawl hosts a remote Streamable HTTP endpoint at https://mcp.firecrawl.dev/{FIRECRAWL_API_KEY}/v2/mcp, with a variant that accepts the key as an Authorization bearer header. Authentication is an API key rather than OAuth; the local stdio package uses the same key via an env var.

Can I point it at a self-hosted Firecrawl instance?

Yes. Set FIRECRAWL_API_URL to your own deployment and the server talks to that instead of the cloud API.

Firecrawl MCP server

OfficialFirecrawl6,500Config last verified Jun 1, 2026

Official Firecrawl server that turns any website into clean, LLM-ready data through scrape, crawl, map, search, and extract.

Firecrawl MCP is the official server from Firecrawl, the web-data API built for AI. It gives an agent a full toolbox for turning the live web into structured, model-ready content: scrape a single URL into clean markdown, batch-scrape a known list of pages, map a domain to discover every indexed URL, crawl a site asynchronously, search the web and optionally scrape the hits, and extract typed structured data with an LLM pass. A research agent tool runs autonomous multi-step gathering, and a set of monitor tools schedule recurring scrapes that diff each snapshot against the last and fire a webhook or email on change.

It installs two ways. Locally it runs over stdio with npx using the firecrawl-mcp package and a FIRECRAWL_API_KEY env var; for self-hosted Firecrawl you can point it at your own instance with FIRECRAWL_API_URL. Firecrawl also hosts a remote Streamable HTTP endpoint at https://mcp.firecrawl.dev/{FIRECRAWL_API_KEY}/v2/mcp, with a bearer-token variant for clients that pass the key in an Authorization header. The API key authenticates requests and meters credits.

Quick install

Copy-paste configs are provided for all 8 supported clients. Pick your client below.

Add to ~/.claude.json

~/.claude.json

json

{
  "mcpServers": {
    "firecrawl": {
      "type": "http",
      "url": "https://mcp.firecrawl.dev/v2/mcp",
      "headers": {
        "Authorization": "Bearer <FIRECRAWL_API_KEY>"
      }
    }
  }
}

Or via CLI

bash

claude mcp add --transport http firecrawl https://mcp.firecrawl.dev/v2/mcp

Step-by-step guides

Available tools

Tool	Description
firecrawl_scrape	Scrape content from a single URL with advanced options, returning clean markdown or other formats.
firecrawl_batch_scrape	Scrape multiple known URLs efficiently with built-in rate limiting and parallel processing.
firecrawl_check_batch_status	Check the progress and retrieve results of a batch scrape operation.
firecrawl_map	Map a website to discover all of its indexed URLs before deciding what to scrape.
firecrawl_search	Search the web and optionally scrape content from the search results.
firecrawl_search_feedback	Submit structured feedback on previous search results to improve quality and refund credits.
firecrawl_crawl	Start an asynchronous crawl job that extracts content from all reachable pages on a site.
firecrawl_check_crawl_status	Check the progress of a crawl job and retrieve results when complete.
firecrawl_extract	Extract structured information from web pages using LLM capabilities against a schema.
firecrawl_agent	Run an autonomous web research agent that browses and gathers data independently and asynchronously.
firecrawl_agent_status	Check the status of an agent research job and retrieve its findings when complete.
firecrawl_monitor_create	Create a recurring page monitor that runs scheduled scrapes or crawls and diffs each snapshot.
firecrawl_monitor_list	List all configured monitors.
firecrawl_monitor_get	Retrieve the details of a single monitor.
firecrawl_monitor_update	Update the settings of an existing monitor.
firecrawl_monitor_run	Trigger an immediate check for a monitor rather than waiting for the schedule.
firecrawl_monitor_checks	List the check history for a monitor with optional filtering.
firecrawl_monitor_check	Get page-level results for a single monitor check, including diffs against the previous snapshot.

Required configuration

FIRECRAWL_API_KEYRequired
Firecrawl API key used to authenticate requests and meter credits. Required for the cloud API.
FIRECRAWL_API_URLOptional
Optional custom endpoint for a self-hosted Firecrawl instance.

What you can do with it

Feed agents clean web content

Scrape a page or search the web and get back tidy markdown the model can reason over directly, skipping a brittle hand-rolled scraping step.

Build a structured dataset from a site

Map a domain to enumerate its URLs, crawl the relevant subtree, then extract typed fields against a schema to populate a database.

Watch pages for changes

Stand up a monitor that re-scrapes on a schedule, diffs each result against the last snapshot, and notifies by webhook or email when something changes.

FAQ

Is it free?: The server is open source under an MIT license and free to run; calls to Firecrawl's cloud API consume credits, and a FIRECRAWL_API_KEY is required. A free tier with starter credits is available, and you can self-host Firecrawl to avoid cloud usage.
Does it support remote/OAuth?: Yes. Firecrawl hosts a remote Streamable HTTP endpoint at https://mcp.firecrawl.dev/{FIRECRAWL_API_KEY}/v2/mcp, with a variant that accepts the key as an Authorization bearer header. Authentication is an API key rather than OAuth; the local stdio package uses the same key via an env var.
Can I point it at a self-hosted Firecrawl instance?: Yes. Set FIRECRAWL_API_URL to your own deployment and the server talks to that instead of the cloud API.

← Browse all search-and-data servers