mcp-local-rag and knowledge-to-action-mcp

These are complements: one provides the RAG retrieval infrastructure (local semantic/keyword search), while the other adds graph-based knowledge organization and agent planning on top, allowing them to work together in an MCP-native knowledge system.

mcp-local-rag
65
Established
knowledge-to-action-mcp
46
Emerging
Maintenance 13/25
Adoption 10/25
Maturity 22/25
Community 20/25
Maintenance 13/25
Adoption 3/25
Maturity 18/25
Community 12/25
Stars: 156
Forks: 32
Downloads: —
Commits (30d): 0
Language: TypeScript
License: MIT
Stars: 3
Forks: 1
Downloads: —
Commits (30d): 0
Language: TypeScript
License: MIT
No risk flags
No risk flags

About mcp-local-rag

shinpr/mcp-local-rag

Local-first RAG server for developers using MCP. Semantic + keyword search for code and technical docs. Fully private, zero setup.

Embeds documents locally using Ollama-compatible models, storing vectors in LanceDB for fast semantic + keyword hybrid search. Integrates with Cursor, Codex, and Claude Code via MCP protocol, while also offering a CLI interface for standalone indexing and querying without requiring an MCP client.

About knowledge-to-action-mcp

tac0de/knowledge-to-action-mcp

MCP server for Obsidian GraphRAG, agent-ready context, preview-only planning, and safe repo handoffs

Combines graph-aware note retrieval with optional embeddings-based semantic reranking to surface contextual neighbors, then structures results into agent-ready packets containing briefs, risks, and repo file hints. Implements preview-only action planning and bounded workspace inspection (ripgrep, git status) without exposing shell access, integrating via stdio with Claude, VS Code, and Cursor.

Related comparisons

mcp-local-rag and mcp-rag-server mcp-local-rag and repo-graphrag-mcp mcp-local-rag and ragmap mcp-local-rag and srag mcp-local-rag and mcp-rag-server mcp-local-rag and Real-time-web-search-RAG-with-MCP

Scores updated daily from GitHub, PyPI, and npm data. How scores work