Claude Code Web Scraping Integration — fastCRW [Firecrawl-Compatible]

Why Claude Code + fastCRW

Claude Code is the agentic coding environment that turns Anthropic's models into a teammate that reads the repo, runs commands, and reasons about web context. The missing primitive is browsing — Claude Code does not ship a built-in scraper. fastCRW closes that gap as an MCP server. Once registered, fastCRW exposes scrape, search, crawl, map, and extract as native Claude Code tools, so any session or Skill can pull live web context. The fastCRW runtime is 6.6 MB of RAM, which makes it sane to run fastCRW locally next to Claude Code instead of routing every request through a heavyweight cloud scraper.

Setup

1. Install the fastCRW MCP server with one command.
2. Set FASTCRW_API_KEY in your shell.
3. Register the server with Claude Code via claude mcp add.
4. Restart Claude Code so it picks up the new MCP tools.

export FASTCRW_API_KEY="fcrw_..."
claude mcp add fastcrw npx -y @fastcrw/mcp

To verify, open Claude Code and run /mcp. You should see fastcrw listed with five tools: scrape, search, crawl, map, and extract.

Code Example

Per-project MCP server registration. Add this to .claude/mcp.json at the repo root so every collaborator gets the same fastCRW tools:

{
  "mcpServers": {
    "fastcrw": {
      "command": "npx",
      "args": ["-y", "@fastcrw/mcp"],
      "env": {
        "FASTCRW_API_KEY": "${FASTCRW_API_KEY}"
      }
    }
  }
}

For self-hosted fastCRW, point the MCP server at your own base URL by setting FASTCRW_BASE_URL:

{
  "mcpServers": {
    "fastcrw": {
      "command": "npx",
      "args": ["-y", "@fastcrw/mcp"],
      "env": {
        "FASTCRW_API_KEY": "${FASTCRW_API_KEY}",
        "FASTCRW_BASE_URL": "https://crw.internal.company.com"
      }
    }
  }
}

Inside a Claude Code session you can now ask the agent to "scrape the latest changelog from anthropic.com using fastCRW" and Claude Code will pick the fastcrw__scrape tool, post the URL to the fastCRW API, and feed the Markdown back into the conversation. The same pattern works for search queries during research tasks.

When to Use This

• Coding sessions that need docs — let Claude Code scrape API references via fastCRW instead of guessing from memory.
• Skills that browse — write a Claude Code Skill that orchestrates fastCRW search plus scrape for grounded answers.
• Slash commands that crawl — define /research to fan out fastCRW search results and scrape the top hits.
• On-call diagnostics — Claude Code can scrape status pages or vendor advisories through fastCRW during incidents.

Limits + Gotchas

• MCP tool calls run inside Claude Code's tool budget. Long fastCRW crawls can exhaust step budgets — prefer scrape and search for in-loop work and run crawls as separate jobs.
• The MCP server reads FASTCRW_API_KEY from the environment. Use a per-machine secret manager rather than committing keys in .claude/mcp.json.
• Claude Code caches MCP tool definitions per session. If you upgrade @fastcrw/mcp, restart Claude Code to refresh the tool list.
• fastCRW responses can be large. Tell Claude Code to summarize before quoting full pages so the context window stays useful.

Related

• MCP integration
• fastCRW MCP docs
• fastCRW scrape docs
• Firecrawl alternative