Collaborative RAG
Provides a multi-tenant Retrieval-Augmented Generation (RAG) server for querying isolated document collections.
About
This Model Context Protocol (MCP) server offers robust Retrieval-Augmented Generation (RAG) capabilities with distinct support for multiple isolated tenants. Users can upload documents in various formats, including PDF, DOCX, and plain text, which are then processed using OpenAI embeddings and FAISS for efficient vector similarity search. Designed for seamless integration with MCP-compatible clients like Claude Desktop and Cline, it can also operate as a standalone server or be accessed directly via its Python API, making it ideal for embedding advanced RAG functionalities into custom applications.
Key Features
- Multi-tenant RAG for isolated knowledge bases
- Flexible document upload via file path or base64 encoding
- Vector search using OpenAI embeddings and FAISS
- Integration with Model Context Protocol (MCP) clients
- 0 GitHub stars
- Supports PDF, DOCX, and plain text document formats
Use Cases
- Developing custom Python applications requiring document querying and knowledge retrieval
- Integrating RAG capabilities into MCP-compatible AI agents like Cline
- Providing a centralized RAG service for multiple teams or applications
Sponsored