SanggonBoy/PyreCrawl

Overview

PyreCrawl is a self-hosted MCP server that gives AI agents web browsing capabilities including scraping, crawling, extraction, site mapping, search, batch operations, research, and monitoring. It requires no API keys and is positioned as a free alternative to Firecrawl. The project is written in Python.

Best for

Best for
Developers who want a free, self-hosted MCP web browsing server for AI agents without API keys.

Use cases

• Extract structured data from web pages for AI agents
• Crawl and map site structures to provide context
• Batch scrape multiple URLs without external API keys

How to use

Install

uvx pyrecrawl@latest

Tools exposed

• PYRECRAWL_CACHE
• PYRECRAWL_CACHE_TTL
• PYRECRAWL_MONITOR_DIR
• PYRECRAWL_NO_TELEMETRY
• Cloudflare-protected
• Self-host

Tested with

Claude Desktop, Claude Code, Cursor, VS Code

Notes

PyreCrawl is a self-hosted MCP server that gives AI agents web browsing capabilities including scraping, crawling, extraction, site mapping, search, batch operations, research, and monitoring. It requires no API keys and is positioned as a free alternative to Firecrawl. The project is written in Python.

12 stars on GitHub. Last updated 2026-09-11. Licensed MIT.

Use cases

• Extract structured data from web pages for AI agents
• Crawl and map site structures to provide context
• Batch scrape multiple URLs without external API keys

Pros

• Self-hosted with no API key requirements
• Covers a broad range of web tasks in one MCP server
• Free alternative to commercial crawling APIs

Cons

• Very early-stage project with only 12 GitHub stars
• Requires self-hosting setup and maintenance
• Python-only, not a managed service

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.