Desktop and iPhone app · 2025–2026
Eisenhower AI
An Eisenhower Matrix task manager that keeps every task as a Markdown file in a folder you own, with AI that sorts a brain dump into tasks you check before they are added.
- My role
- Independent developer: product, design and engineering
- Status
- Out on Windows and Mac; iPhone and iPad in development
- Platforms
- Windows: Version 0.3.1, also in the Microsoft Store
- Mac: Version 0.2.0, signed and notarised by Apple
- iPhone and iPad: In development
Free download
Get Eisenhower AI
Or get it from the Microsoft Store- Windows: version 0.3.1, released October 14, 2026. Windows 10 or 11, 64-bit.
- Mac: version 0.2.0, released October 13, 2026. macOS 13 Ventura or later, on Apple silicon or Intel.
- The iPhone and iPad app is in development.
New on Windows since 0.2.0
- The app updates itself: when a new version is out, it asks once and installs it.
- Voice input: speak, and your words are turned into text on your own computer.
- In the Microsoft Store, which signs the app and keeps it up to date.
- The AI Classifier explains itself the first time, with an example to try.
- Simpler Settings and a two-step welcome tour.
Installing
Windows
The Microsoft Store version installs like any Store app and the Store keeps it up to date. The direct installer needs no administrator rights, but it is not signed yet, so Windows may show “Windows protected your PC”: choose More info, then Run anyway. From then on the app updates itself.
Mac
Open the downloaded file and drag Eisenhower AI into Applications. The app is signed and notarised by Apple, so it opens like any other.
Privacy
- Your tasks stay in a folder on your computer. You don't need an account, and nothing is tracked.
- Task text leaves your computer only when you use AI: to LM Studio on your own computer, or to the cloud provider whose API key you added.
- API keys are kept in the Mac Keychain or Windows Credential Manager.
Checksums (SHA-256)
- Eisenhower-AI-0.2.0-mac.dmg
- efa4af391d0c06813d653a71a5532781f035a3a2ec29f14550f201b37554727f
- Eisenhower-AI-0.3.1-windows-setup.exe
- e0eedafb816f285256afcb9b4e370b59528141ea1d5b89c82f5da73c5cdccd44
01 · The idea
A serious local-first productivity surface
The project starts from the Eisenhower Matrix but treats it as a real product: task state, reminders, recurring work, editing, storage, backups, import/export, and desktop ergonomics.
02 · How it works
Quadrants are only the visible layer
The useful engineering work lives beneath the four boxes: a shared task core, a Markdown store that the user can read and back up, reminders, preferences, and an AI step that only ever suggests.
- Tasks are Markdown files in a folder the user owns; the app keeps no other copy.
- Desktop-first speed without forcing cloud sync or an account.
- AI suggestions that stay explainable, reviewable and reversible.
03 · The hard part
Make automation feel controllable
The AI Classifier splits a pasted paragraph into tasks, suggests a quadrant and picks up dates, then shows every suggestion for review. If a provider is unavailable the app says so and keeps the text for a retry instead of guessing.
04 · What's next
iPhone and iPad
The iPhone and iPad app reads and writes the same Markdown task format. Its AI Capture turns a quick brain dump into sorted tasks, and the same Pro account works on the desktop.
~/contributions
What I did
- Designed and built the original Python/PyQt desktop application, including quadrant interactions, persistence, settings, import/export, statistics, and Windows packaging.
- Rebuilt the product as a TypeScript monorepo with a React/Tauri desktop app, shared task rules, and a storage package that keeps each task as a Markdown file in a user-owned folder.
- Added optional AI classification through LM Studio on the user's own computer or the user's own cloud-provider key, with review before anything is added and separate evaluation scripts.
- Building the iPhone and iPad app in Expo and React Native on the same Markdown task format, with AI Capture behind Sign in with Apple and a Pro subscription.
~/outcomes
Results
- Released for Windows and Mac and listed in the Microsoft Store; on Windows the app updates itself from signed update files.
- Keeps every task as a readable Markdown file in a folder the user chooses, with undo, a trash, daily backups and CSV or JSON export.
- AI sorting is optional and reviewed: a model suggests tasks and quadrants, and nothing is added until the user confirms.
- An English and a Turkish interface, light, dark and high-contrast themes, reminders, recurring tasks and voice input.
Key decision
How much authority should AI have over a user's priorities?
Choice
Let the model propose tasks, quadrants and dates, then keep the user in control of confirmation and final placement.
Why
Priority is a commitment, not only a text-classification problem. A review step makes the suggestion explainable, reversible, and compatible with a local-first workflow.
Tradeoff
Confirmation adds a small interaction cost, but avoids silent task movement and preserves trust when a model is uncertain.
~/stack
Built with
- TypeScript
- React
- Tauri
- Rust
- Expo / React Native
- Supabase
- LM Studio
~/verification
Where to check it
- Public releases
Windows installers, SHA-256 checksums and signed update files are published for each version, and the app is listed in the Microsoft Store.
- Original application source
The public Python/PyQt repository documents SQLite storage, drag-and-drop, statistics, JSON/CSV transfer, system-tray behavior, and standalone Windows packaging.
Release checks
Each release runs a verification gate first: JavaScript and Rust tests, the workspace build, formatting checks and a scan for secrets.
- Legal pages
Privacy policy, terms of use and support pages for the app are published in their own public repository.
More screenshots
