Built on LlamaIndex with pgVector, this server connects Claude to your self-hosted RagRabbit instance for semantic search across your website content. It exposes search operations that let Claude retrieve relevant pages from your documentation using vector similarity. You'd reach for this when you want Claude to have contextual knowledge of your specific site content without manual copy-pasting. The server requires a deployed RagRabbit instance (one-click Vercel deploy available) and works with both Claude Desktop and Cursor IDE. It's essentially a bridge between Claude and your crawled, indexed website content stored in PostgreSQL.
claude mcp add --transport stdio madarco-ragrabbit uvx ragrabbit