Librecrawl Technical Seo Audit Mcp
@adityaarsharma
Open-source self-hosted technical SEO audit MCP server built on the LibreCrawl engine. Runs full audits inside Claude, Cursor, or Codex — 37 tools, 50+ checks per audit (hreflang, schema.org Rich Results, security headers, image performance, WAF detection on 200-OK pages), chunke
概览
What is librecrawl-technical-seo-audit-mcp?
librecrawl-technical-seo-audit-mcp is an AI-native technical SEO crawler that runs inside Model Context Protocol (MCP) clients like Claude, Cursor, and Codex. It performs unlimited-page on-site SEO audits with 50+ checks, produces branded PDF reports and CSVs, and is MIT-licensed, self-hosted, and ephemeral by design.
How to use librecrawl-technical-seo-audit-mcp?
Install with the provided curl command or manually using Python and Docker. The installer configures your MCP client. To use it, simply ask your AI assistant to audit a website — the assistant calls the server’s 37 MCP tools to run the crawl and return results.
Key features of librecrawl-technical-seo-audit-mcp
- Unlimited pages with chunked-progressive crawling (no timeout)
- 50+ technical SEO checks including hreflang, schema, and redirects
- WAF/bot-block detection (Cloudflare, Akamai, DataDome, Imperva, PerimeterX)
- Auto-adaptive crawl delay using AIMD algorithm
- Ephemeral by design — session data deleted after download
- Branded PDF reports ready for client delivery
- Google Search Console merge support
Use cases of librecrawl-technical-seo-audit-mcp
- Full technical SEO audit for any website, from small blogs to 10,000-page enterprise sites
- Identifying broken links, redirect chains, and orphan pages
- Checking hreflang, canonical, and schema markup compliance
- Generating branded audit reports for client handoffs
- Bulk auditing multiple client sites without persistent data
FAQ from librecrawl-technical-seo-audit-mcp
Is there a page cap or paid tier?
No. The server is free, MIT-licensed, and self-hosted with no page limit.
What are the runtime requirements?
Python 3.10+ and Docker (for the LibreCrawl backend) are required.
Does it detect WAF challenges that return 200 OK?
Yes. It fingerprints challenges from Cloudflare, Akamai, DataDome, Imperva, and PerimeterX in the response body and flags them.
Does the server persist audit data after I download the results?
No. Once you download the zip, the server deletes the session row, all artifact files, and the upstream crawl record — zero bytes remain.
What reports does it produce?
A branded PDF report and 7 CSV files. With GSC merge, an additional 4 CSVs.
