nkapila6/mcp-local-rag
"primitive" RAG-like web search model context protocol (MCP) server that runs locally. ✨ no APIs ✨
Implements multi-engine web search across 9+ backends (DuckDuckGo, Google, Bing, Brave, Wikipedia) with semantic ranking using Google's MediaPipe text embeddings, extracting markdown from fetched URLs without external APIs. Exposes tools like `deep_research`, `deep_research_google`, and `rag_search_ddgs` as MCP resources compatible with Claude Desktop, Cursor, and other MCP clients. Deployable via `uvx` or Docker and includes Agent Skills that guide LLMs on query formulation and backend selection for privacy-aware or comprehensive research.
118 stars.
Stars
118
Forks
19
Language
Python
License
MIT
Category
Last pushed
Mar 09, 2026
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/mcp/nkapila6/mcp-local-rag"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
shinpr/mcp-local-rag
Local-first RAG server for developers using MCP. Semantic + keyword search for code and...
AmberLee2427/nancy-brain
Nancy's RAG backend and HTTP API/MCP server connectors.
tac0de/knowledge-to-action-mcp
MCP server for Obsidian GraphRAG, agent-ready context, preview-only planning, and safe repo handoffs
HaseebKhalid1507/VelociRAG
Lightning-fast RAG for AI agents. ONNX-powered, 4-layer fusion, MCP server. No PyTorch.
RoboFinSystems/robosystems
RoboSystems is a financial knowledge graph platform that transforms complex financial data into...