explaingit

reorproject/reor

8,561JavaScriptAudience · pm founderComplexity · 2/5LicenseSetup · moderate

TLDR

A desktop note-taking app that runs AI entirely on your own computer to find connections between notes and answer questions about your writing, with no data sent to the cloud.

Mindmap

mindmap
  root((Reor))
    Core features
      Markdown editor
      Related notes
      AI chat search
      Local storage
    AI models
      Ollama local
      OpenAI compatible
      Download in app
    Privacy
      No cloud sync
      All data local
      AGPL license
    Setup
      Mac Linux Windows
      Choose folder
      Import markdown
Click or tap to explore — scroll the page freely

Code map

Detail Auto

An interactive map of this repo's files and how they connect — its source is parsed live in your browser. Click Visualize to build it.

filefunction / class

Things people build with this

USE CASE 1

Ask questions about your entire notes collection in plain language and get answers sourced from your own writing.

USE CASE 2

Automatically surface related notes as you write, based on meaning rather than just keyword matching.

USE CASE 3

Run AI-powered note search offline on your laptop with no internet connection or subscription fees.

USE CASE 4

Connect to a local Ollama model or any OpenAI-compatible API to power the AI features privately.

Tech stack

JavaScriptElectronOllama

Getting it running

Difficulty · moderate Time to first run · 30min

Requires installing Ollama separately and downloading a local AI model before AI features are available.

Free to use and modify, but if you distribute a modified version you must release your changes under the same AGPL-3.0 license.

In plain English

Reor is a desktop note-taking app that runs AI tools entirely on your own computer, with no data sent to external servers. It works on Mac, Linux, and Windows. You write notes in a markdown editor that looks similar to Obsidian, and everything is stored in a folder you choose on your filesystem. The core idea is that your notes are not just stored but actively processed. Each note is split into chunks and turned into a numeric representation stored in a local database. This lets the app find connections between notes automatically, based on what they are actually about rather than just keywords. When you open a note, a sidebar can show you related notes pulled from your collection. Reor also lets you ask questions about your notes in a chat-style interface. Instead of searching manually, you type a question and the AI looks through your notes to find relevant passages and uses them to form an answer. This is useful when you have a large collection and want to find what you wrote about a topic without remembering the exact file or phrase. The AI models that power all of this run locally through Ollama. Inside the app you can download and switch between models from Ollama's library. You can also connect to other OpenAI-compatible APIs if you prefer. Importing notes from another app means copying your markdown files into the chosen folder manually, since there are no direct integrations yet. The project is open-source under AGPL-3.0 and welcomes contributors. The name comes from the Latin word for to think.

Copy-paste prompts

Prompt 1
I want to move my Obsidian vault into Reor, what do I copy and where, and will my existing markdown files work without changes?
Prompt 2
Set up Reor with a local Ollama model on my Mac so I can ask questions about my notes without sending data anywhere.
Prompt 3
How does Reor find related notes, explain how the local embeddings and vector database work so I can tune the experience.
Prompt 4
Connect Reor to an OpenAI-compatible API endpoint instead of Ollama, show me which settings to change.
Open on GitHub → Explain another repo

← reorproject on gitmyhub — every repo by this author, as a profile.

Verify against the repo before relying on details.