What's already configured
The one-click install brings up Firecrawl's Docker Compose stack automatically: the API service, a Redis instance for queueing, a Postgres-backed task queue (NuQ), and a Playwright service that handles JavaScript rendering for scraping targets. The API is published on port 3002 and reachable at http://your-server-ip:3002 once the containers finish starting. By default, self-hosted Firecrawl runs with USE_DB_AUTHENTICATION=false — meaning the API accepts requests without an API key out of the box. This is fine for a private network or initial testing, but if you plan to expose the endpoint publicly or share it across a team,.
Making your first request
Firecrawl doesn't ship a web login panel — it's an API-first tool, so "logging in" means sending your first authenticated (or unauthenticated, by default) POST request to an endpoint like /v1/scrape with a target URL in the body. A basic health check is confirming the API responds at http://your-server-ip:3002 before wiring it into any client. If you plan to connect Firecrawl to an AI agent or IDE, it also ships an MCP server and CLI that point at this same API URL — useful if you're integrating with tools like Claude, LangChain, or n8n rather than calling the REST API directly.
Before you go further
Two values matter for security once the instance is live: the queue authentication key and the Postgres password, both set via environment variables in the stack's .env configuration and left at documented defaults after a one-click install.. Rotate both before pointing production scraping jobs — especially anything triggered by external webhooks — at the instance.