Optional: Add to Claude
- Open the Firecrawl connector using the button above.
- Connect your Firecrawl account and approve access when prompted.
- Start a new Claude conversation, enable Firecrawl, and ask it to search or read a page.
Optional: Set up Claude Code
Connect Claude Code through the hosted MCP server. Run the command below, then sign in through your browser.1. Add Firecrawl
Run:https://mcp.firecrawl.dev/v2/mcp-oauth is a client configuration value, not a page to open directly in a browser.
2. Sign in and verify
Enter/mcp in Claude Code, choose firecrawl, and complete the browser sign-in. Start a new Claude Code session if you changed an existing Firecrawl connection.
Confirm that only one firecrawl server is active, then try:
What you can do
Search
Find current information and read the content of relevant results.
Scrape
Turn URLs into clean markdown or structured data.
Interact
Navigate, click, and fill forms to reach the data you need.
Alexandria
Discover structured data tools, inspect inputs and pricing, then run them through Firecrawl.
Other connection modes
- Try keyless MCP for Search, Scrape, and Parse without an account or key.
- Configure an API key for unattended use. Keep the secret outside agent chat and use an environment-variable reference.
- Run MCP locally only when you need a local process or a self-hosted Firecrawl API. The local package requires Node.js 22 or newer.
Optional: literature research
For paper and literature work, the research-index skill routes those queries to thefirecrawl_research_* tools and the Research Index instead of general web search:
Troubleshooting
- A Firecrawl server already exists: run
claude mcp remove firecrawl, add the intended connection once, then start a new session. - The browser does not open: enter
/mcp, selectfirecrawl, and start authentication from the client. - The server shows
401: sign in again through/mcp; do not paste an API key into the conversation or OAuth URL.

