Show & Tell: librecrawl-mcp — SEO audit MCP server for Claude / Cursor / Codex (built on LibreCrawl)

[adityaarsharma]

Hi Phiality 👋

First thank you. LibreCrawl is the engine I've been waiting for: free,
unlimited pages, no SaaS lock-in, real technical-SEO depth. I wanted that
power inside my AI assistant, so I built a Model Context Protocol (MCP)
wrapper around it:

→ https://github.com/adityaarsharma/librecrawl-mcp (MIT)

What it adds on top of LibreCrawl (your engine does ALL the actual crawling):

• 37 MCP tools so Claude Code / Claude Desktop / Cursor / Codex / Windsurf
  can drive a full audit by chat
• Chunked-progressive runner (background worker + SQLite WAL) so MCP clients
  don't time out on big sites
• AIMD-adaptive crawl-delay controller
• WAF / bot-block fingerprinting (Cloudflare / Akamai / DataDome / Imperva
  / PerimeterX) on 200-OK responses
• 50+ extended checks on top of your existing data (hreflang, schema.org
  Rich Results, security headers, content quality, etc.)
• Branded PDF report + 7 sidecar CSVs (WeasyPrint)
• Ephemeral-by-default: zip returned to client, server wipes everything

LibreCrawl is credited prominently in the README (badge + hero line + Credits)
and the project clearly positions itself as a WRAPPER, not a replacement.
Code is MIT, matching upstream.

A few asks if you're interested:

1. Would you be open to a one-line link in LibreCrawl's README pointing to
   librecrawl-mcp under something like "Use LibreCrawl from your AI
   assistant"? Happy to send the PR.
2. Any concerns with the name librecrawl-mcp? Picked it specifically to
   credit upstream, not to imply ownership — happy to rename if it
   bothers you.
3. If you ever want a contribution back (the WAF fingerprinting code, the
   schema validator, anything), let me know — I'd be glad to upstream it.

No expectation either way. Mainly wanted you to know it exists and to say
thanks for the work.

— Aditya (adityaarsharma.com)

[pedrovillalobos]

This would look great on LibreCrawl. Why don't you just imlpement your suggestions and make a pull request? :)