Browse
→ Productivity
→ Obsidian Vault Search MCP
Obsidian Vault Search MCP
Local-first Obsidian search MCP server using Python standard-library code, SQLite indexing, BM25 retrieval and optional OpenAI-compatible semantic reranking or aggregation. It keeps vault indexes on disk, supports Chinese bigram tokenization and works without an API key for basic retrieval. Public GitHub repository created August 25, 2026.
MCP unverified
Integration
| Transport | stdio |
| Auth | none |
| Endpoint | https://github.com/johnadams-bot/obsidian-vault-search-mcp |
| Install | git clone https://github.com/johnadams-bot/obsidian-vault-search-mcp.git |
Use Cases
| 01 | Find Obsidian notes by meaning or fuzzy remembered content from an MCP client |
| 02 | Build a local SQLite index of a vault without sending notes to a hosted service |
| 03 | Optionally rerank or summarize matching notes with any OpenAI-compatible LLM endpoint |
Tags
obsidian notes local-search bm25 sqlite chinese knowledge-base python
Machine-readable: /api/servers.json
· JSON-LD schema embedded in <head>