obsidian-notes-rag
An MCP server and CLI for semantic search over an Obsidian vault or any folder of linked Markdown files, generating embeddings through OpenAI, Ollama or LM Studio and storing vectors locally in sqlite-vec with no telemetry.
obsidian-notes-rag indexes a folder of Markdown files - an Obsidian vault, a documentation tree, a wiki export - and lets a user or an AI assistant search it by meaning rather than keyword match. A setup wizard writes a configuration file with the vault path, embedding provider and API key, and an index command chunks files by heading structure using Chonkie's RecursiveChunker, generates embeddings through OpenAI, Ollama or LM Studio, and stores everything in a local SQLite database using sqlite-vec, with no telemetry or unnecessary network calls.
The CLI, installed as both obsidian-rag and okf-search, supports search, similar and context commands from the terminal, and the same functionality is exposed as an MCP server so any compatible client - Claude Code, Claude Desktop, Cursor, Windsurf - can search, find related notes and pull context during a conversation. Nothing in the tool requires Obsidian itself; pointing --root at any Markdown directory works the same way.
An AI coding assistant can also call the installed CLI directly through shell commands instead of loading the MCP server, avoiding extra tool definitions in its context window.