Lesson 2 — Getting Started with Cloud Services

⏱ Est. reading time: 4 min Updated on 5/7/2026

2.1 Obtaining an API Key

  1. Visit firecrawl.dev and click "Get started."
  2. Register with GitHub or Email (no credit card required for the free plan).
  3. Go to Dashboard → API Keys → Create a new Key.
  4. Copy the key in the fc-xxxxxxxxxxxx format.

Free Credits: New users receive 500 credits (one-time), which is enough for comprehensive testing and evaluation.


2.2 Configuring in Claude Code

Edit ~/.claude/settings.json (global) or .claude/settings.json in your project root:

{
  "mcpServers": {
    "firecrawl": {
      "command": "npx",
      "args": ["-y", "@anthropic/firecrawl-mcp"],
      "env": {
        "FIRECRAWL_API_KEY": "fc-your-api-key"
      }
    }
  }
}

After configuring, restart Claude Code, and the MCP tools will load automatically.


2.3 Configuring in Cursor

  1. Open Cursor Settings (Cmd + ,).
  2. Navigate to FeaturesMCP Servers.
  3. Click "+ Add new global MCP server."
  4. Enter the following:
    • Name: firecrawl
    • Type: command
    • Command: npx -y @anthropic/firecrawl-mcp
    • Environment Variables: FIRECRAWL_API_KEY=fc-your-api-key

2.4 Your First Scrape Request

In Claude Code or Cursor, simply type:

"Use firecrawl_scrape to fetch the content of https://docs.firecrawl.dev in markdown format."

Result: You will get clean Markdown content. Firecrawl automatically removes navigation, ads, and footers.


2.5 Your First Search Request

The Search tool not only searches but also fetches the content of the search result pages:

"Search for 'Firecrawl MCP setup guide' using firecrawl_search and return the markdown content of the top 3 results."

MCP Call Example:

{
  "tool": "firecrawl_search",
  "arguments": {
    "query": "Firecrawl MCP setup guide",
    "limit": 3,
    "scrapeOptions": { "formats": ["markdown"] }
  }
}

2.6 Verifying MCP Status

Type the /mcp command in Claude Code. If you see firecrawl with a status of connected, you are all set!