A semantic documentation search server that connects Claude to a Qdrant vector database for RAG operations. Exposes three core tools: add_documentation for ingesting docs from URLs, search_documentation for natural language queries, and list_sources for inventory management. Uses Ollama's nomic-embed-text model by default or OpenAI embeddings as an alternative. You'd reach for this when you need Claude to search through large documentation sets without hitting context limits. Works with both local Docker-based Qdrant instances and Qdrant Cloud. Integrates with Cline, Roo-Code, and Claude Desktop through standard MCP configuration.
claude mcp add --transport stdio qpd-v-mcp-ragdocs uvx mcp-ragdocs