Qurio is an open-source, self-hosted ingestion and retrieval engine designed to function as a local 'Shared Library' for AI agents, including tools like Cursor, Windsurf, or custom scripts. Unlike cloud-based RAG solutions, Qurio runs entirely on your machine, ingesting handpicked heterogeneous documentation such as web crawls, PDFs, and Markdown files. It serves this curated context directly to your IDE via the Model Context Protocol (MCP), enabling AI to generate more accurate code faster by relying on trusted information. Qurio emphasizes privacy by keeping your data local, enhances precision to prevent AI hallucinations, offers rapid deployment via Docker Compose, and is built on open standards like MCP, Weaviate, and PostgreSQL.
主要功能
01Universal Ingestion: Crawl documentation sites or upload files (PDF, DOCX, MD).
02Configurable Reranking: Integrate Jina AI or Cohere for precision tuning.
03Native MCP Support: Exposes a standard JSON-RPC 2.0 endpoint for seamless integration with AI coding assistants.
04OCR Pipeline: Automatically extracts text from scanned PDFs and images via Docling.
05Hybrid Search: Configurable BM25 keyword search with Vector embeddings for high-recall retrieval.
060 GitHub stars
