Plugin: Similarity — Find Notes by Meaning (Fully Local)

Share & showcase
1

Similarity — Find Notes by Meaning (Fully Local)

I started this over a year ago, but here’s the stable release; a plugin that helps you find notes based on meaning, not just keywords.

This plugin’s priorities:

  • Fully local, by design – your data never leaves your device. No APIs, no background calls, no risk.

  • Lightweight and fast – minimal overhead, built to stay responsive even as your vault grows.

  • Keeping it simple – hiding all the complicated parts so you can focus on writing.

  • Focused experience – no fancy AI LLM agents, just practical semantic discovery that fits naturally into your workflow.

  • Cross-platform consistency – works on desktop and mobile, with a shared index so you don’t need to reindex per device.

What it does

Browse related notes

Jump through notes that are actually related in meaning—not just linked.

leaf-demo

Search by meaning

Look up notes based on what you’re trying to say, not how you phrased it.

modal-demo

Why I built this

I found myself:

  • forgetting notes existed

  • over-organizing with folders/tags

  • missing connections across topics

Embeddings solve this surprisingly well—but most tools around it feel heavy or require setup.

I wanted something that just works.

If you journal in Obsidian, I would highly recommend trying this plugin. It’s been very insightful to find journal entries of similar times in my life.

Install (via BRAT)

Still waiting on official approval, but you can install it via BRAT:

https://github.com/JoramMillenaar/obsidian-related-notes

Curious how others are using semantic search in their vault—open to feedback or ideas.

P.S. giving the repo a star would be appreciated :star:! It’ll help visibility with the Obsidian team :).

2 Likes
2

Have you tried Graph Analysis?
If so, how does your plugin compare?

(I don’t like to use LLMs with RAG for semantic searches, either, because I get too many irrelevant results.
Instead, I use AI IDE’s with natural language questions to target topics on demand.)

3

Graph Analysis looks powerful—I’ll check it out. From what I can tell, it focuses more on structural relationships (links, co-citations, shared neighbors), while Similarity relates notes based on actual content.

For me, that shifted things a lot—I barely do manual linking anymore. Seeing what’s similar in meaning covers most of that.

Yeah, I’m with you on RAG/LLMs too. They’re powerful, but less deterministic, and I’m not a fan of sending my notes to a third party :upside_down_face: .

This is plugin is kind of a RAG system, I’m just the LLM

1 Like
4

Also here’s the Repo link that I forgot to add; Similarity.

5

Graph Analysis is quite powerful, but also crude. This looks worth trying out.

1 Like