jaimenbell/rag-mcp

Overview

A Model Context Protocol server for retrieval-augmented generation over a local corpus. Its searchknowledge tool returns cited passages with source, heading, and chunk index. Runs fully local with zero-cost embeddings, stores vectors in ChromaDB, and supports incremental ingestion.

Best for

Best for
Developers building local RAG workflows on MCP with minimal cost

Use cases

• Add local RAG queries to an MCP-compatible client
• Search a document corpus and get cited excerpts
• Incrementally index new documents into ChromaDB

Notes

A Model Context Protocol server for retrieval-augmented generation over a local corpus. Its searchknowledge tool returns cited passages with source, heading, and chunk index. Runs fully local with zero-cost embeddings, stores vectors in ChromaDB, and supports incremental ingestion.

0 stars on GitHub. Last updated 2026-08-21. Licensed MIT.

Use cases

• Add local RAG queries to an MCP-compatible client
• Search a document corpus and get cited excerpts
• Incrementally index new documents into ChromaDB

Pros

• Fully local and private, no external API calls
• Zero embedding cost
• Results include source and chunk citations

Cons

• No community traction or production track record yet
• Requires ChromaDB setup and management
• Limited to corpora you have already ingested

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.