crw — Fast Web Scraping + Search MCP in Rust

Main

• Use local mode first to validate output quality and latency before you depend on it in a larger agent workflow.

• If you need API compatibility, keep the Firecrawl-style endpoint paths in your client code so switching providers stays easy.

• Treat rendering as an operational variable: allow fallback renderers and surface warnings to your agent logs.

• For production, add rate limiting and host allowlists so web access doesn’t become an unbounded crawler.

README (excerpt)

The web scraper built for AI agents. Single binary. Zero config.

Works with: Claude Code · Cursor · Windsurf · Cline · Copilot · Continue.dev · Codex · Gemini CLI

Quick Start • AI Agents •

Source-backed notes

• README includes a quick start install script via curl ... install.sh | CRW_BINARY=crw sh followed by crw setup.
• README states it provides a Firecrawl-compatible API and includes MCP integration examples for Claude Code via npx crw-mcp.
• README contains benchmark tables and resource claims; treat them as README-provided context, not independently verified metrics.

FAQ

• Is it only an API server?: README describes both a CLI and MCP integration; you can start with the CLI locally.
• How do I connect it to my agent?: README shows adding it to Claude Code via claude mcp add ... -- npx crw-mcp.
• How do I keep web access safe?: Use allowlists, rate limits, and start with minimal scope in local mode.