GitHub Repository

Semantic search for Obsidian vaults using LanceDB and cloud or local embedding models

14 starsPython

Obsidian-Semantic, a CLI that lets agents search your vault by meaning

Name: Obsidian-Semantic, a CLI that lets agents search your vault by meaning
Availability: InStock
Author: ravila4

by ravila4·May 8, 2026·4 points·2 comments

Visit Project View on HN

AI Analysis

●●SolidNiche GemSolve My Problem

Smart Connections plugin but CLI-first with local Ollama support.

Strengths

•Incremental indexing saves time compared to full re-indexes on large vaults.
•Section-level retrieval via heading paths enables precise context injection.
•Deduplication logic prevents flooding agents with redundant note chunks.

Weaknesses

•Score thresholds vary wildly between embedders, requiring manual calibration.
•Lacks a GUI for non-technical users who prefer visual graph exploration.

Post Description

Hi HN, I built this for myself because I wanted my coding agent (Claude Code) to actually be able to use my Obsidian vault as a knowledge base, not just grep it.

The use I get the most mileage from is asking the agent to find notes that should be cross-linked, which surfaces forgotten connections and turns the vault into more of a wiki over time.

It is similar to what the Smart Connections Obsidian plugin does, but I wanted a CLI-first tool, and more control over the models. Currently it supports local embedding models via Ollama and LMStudio, plus Gemini API for cloud.