SimpleNotes

A menu bar app that turns freeform notes into actionable checklists using a local AI model (Ollama + Qwen). Lives in your menu bar, no cloud required. Works on macOS and Windows.

Prerequisites

• macOS (Apple Silicon or Intel) or Windows 10/11
• Node.js v18 or later — download here (only needed for development)
• 16 GB RAM recommended (the Qwen 3 8B model uses ~8 GB)

Quick Start

From a packaged build (recommended)

1. Download the latest release for your platform (.dmg for macOS, .exe installer for Windows)
2. Open the app — a setup wizard will walk you through installing Ollama and downloading the AI model
3. Start using it

From source

git clone https://github.com/isaacHuh/simple_notes.git
cd simple_notes
npm install
npm start

On first launch, the setup wizard will guide you through any remaining setup (installing Ollama, pulling the model). You can also set things up manually — see below.

Setup Wizard

When you first open SimpleNotes (or if Ollama/the model becomes unavailable), a guided setup wizard appears:

Step 1 — Ollama installed & running

• The app checks if the Ollama server is reachable at localhost:11434
• If not found, click "Download Ollama" to open the correct download page for your platform
• After installing, open the Ollama app and click "Check again"

Step 2 — Model ready

• The app checks if the configured model (default: qwen3:8b) is downloaded
• If not, click "Pull Model" to download it directly from the app with a live progress bar
• The download is ~4.7 GB

Once both steps show green checkmarks, click "Get Started".

You can also click "Skip setup" if you prefer to configure things manually or use a different model.

Manual Setup (alternative)

If you prefer to set up Ollama yourself instead of using the wizard:

Install Ollama

macOS:

brew install ollama

Or download from ollama.com/download/mac.

Windows: Download the installer from ollama.com/download/windows.

Start the Ollama server

macOS: Open the Ollama app, or run in a terminal:

ollama serve

Windows: Open Ollama from the Start menu. The server starts automatically.

The server runs on http://localhost:11434.

Pull the model

ollama pull qwen3:8b

Verify it works

ollama run qwen3:8b "List 3 things to do before a road trip"

Usage

1. Type a note in the input bar at the bottom — anything like "Meeting notes: discuss Q3 budget, review hiring pipeline, schedule team offsite"
2. Press Enter — the note is sent to your local model, which extracts actionable checklist items
3. Check off items as you complete them — they move to the "Completed" section
4. Drag and drop one task onto another to merge them with AI
5. Click + on any task to add context or sub-tasks
6. Right-click any item to delete it
7. Green dot in the title bar = Ollama is connected. Red dot = Ollama is not reachable.

Your checklist is saved automatically and persists across restarts.

• macOS: ~/Library/Application Support/NoteFlow/data.json
• Windows: %APPDATA%/NoteFlow/data.json

Troubleshooting

Red status dot / "Ollama is not running"

• Make sure Ollama is open (macOS: check menu bar; Windows: check system tray)
• Or run ollama serve in a terminal
• Check that nothing else is using port 11434

"Model not found"

• Run ollama pull qwen3:8b to download the model
• Or use the setup wizard (it will appear automatically)
• Verify with ollama list

Slow responses

• Close memory-heavy apps (browsers with many tabs, Docker, etc.) to free up RAM
• Check Activity Monitor (macOS) or Task Manager (Windows) — memory usage should have headroom

App doesn't appear in menu bar / system tray

• macOS: Look for the icon in the top-right area of your menu bar. On Macs with a notch, try Cmd+drag to rearrange icons
• Windows: Check the system tray (bottom-right, click the ^ arrow if hidden)

Build for Distribution

macOS (produces .dmg and .zip):

npm run build

Windows (produces NSIS installer and .zip):

npm run build:win

Both platforms:

npm run build:all

Output goes to the dist/ directory.

Tech Stack

• Electron + menubar — tray icon and floating panel
• Ollama — local AI inference server
• Qwen 3 8B — language model for task extraction
• electron-builder — packaging for macOS (DMG) and Windows (NSIS)
• Vanilla HTML/CSS/JS — no frontend framework

Contributing

Contributions are welcome! Please see CONTRIBUTING.md for guidelines.

License

This project is licensed under the MIT License.