Ayos: Photo Cleaner is now available on iOS and Android.
Ayos is on the App Store and Google Play. It finds the photos that were never meant to be memories, explains each call, and does all of it on the device.
Ayos: Photo Cleaner is now public on both platforms: on the App Store for iPhone and on Google Play for Android. Both are free to download. The product page is at horizondigital.dev/ayos.
What it does
Every camera roll has a junk drawer: the receipt photographed for an expense report, the parking-garage pillar, the Wi-Fi password on a whiteboard, the serial number on the back of a router, the shelf photographed to remember what to buy. None of these were meant to be kept, and none of them get deleted, because finding them by hand means scrolling past everything that was.
Ayos finds them. It sorts a library into piles: Possible Not a Memory (receipts and documents, products and shopping, food, memes and saved images, utility shots, old screenshots), exact duplicates with the best copy kept, similar shots of the same moment, screenshots and screen recordings, blurry photos, and large videos. Every photo in a pile carries a plain-English reason: This looks like a receipt from March 2023. Still need it?
Ayos suggests; the person decides. High-confidence items come pre-checked, everything else is one tap away, and the uncertain ones can be reviewed one at a time. The deleting is done by the phone's own photos app, not by Ayos.
Everything stays on the device
Analysis runs on the phone. On iPhone that is Apple's on-device frameworks; on Android it is Google's on-device ML Kit, with the models bundled into the app. There is no account, no cloud, no analytics, and the app makes no network connections of its own. Nothing about a library, not the photos, not the metadata, not the analysis, ever reaches us. The privacy policy says so in plain words, and the App Store listing's privacy label reads "data not collected" because that is the truth of it.
That was the constraint the app was designed under, not a feature added afterwards. Most of our work is about protecting the data people keep. Ayos is the other half: deciding what not to keep, with the decision explained and left to the person. It is also a working test of how far on-device image analysis can go without sending a single photo anywhere, built twice on two different frameworks. A writeup of what that taught us will follow once the app has been in real hands for a while.
Deleting is slow on purpose
Nothing is deleted until one button is pressed. After that, the chosen photos wait inside Ayos for seven days, where a tap takes any of them back, before they go to the photos app. The photos app asks once more, then keeps them in its own trash for thirty days. Anyone in a hurry can press Delete now; nobody loses a photo to a stray tap. Favorites are never suggested, and photos of people, pets, and trips rank last for any suggestion.
Two platforms, one product
The Android app is a native port of the iPhone app, not a wrapper around it: the same piles, the same rules about what is protected, the same reason on every photo, and the same seven-day stage before anything leaves. Both are free to download with a limited number of free deletions; an optional Pro purchase, made through the store, removes the limit. A Mac version follows.
Where to get it
- Ayos: Photo Cleaner on the App Store
- Ayos: Photo Cleaner on Google Play
- Product page and the privacy policy
- Support: hello@horizondigital.dev, answered by the person who wrote the app
If you work on this too
These are written for people already deep in the same problems. Corrections, disagreement, and better approaches are all welcome — particularly the disagreement.