How we research

Who writes these

aisoftware, the developer of PhotoBroom. We are not a review site and we are not neutral. What we can do is show our sources and name the cases where a competitor is the better buy — and there are several, each called out on the page it belongs to rather than buried here.

What we know first-hand

One app on this site is not researched. PhotoBroom is ours — we write it, ship it, and run it against our own photo libraries on real devices. So when a page says what PhotoBroom does, that is not a reading of a listing. It is the behaviour of software we built, and we can be held to it.

That means the two halves of every comparison rest on different foundations, and it is better to name that than to blur it. Everything about PhotoBroom here is first-hand. Everything about the app it is compared with is documentary — read from that vendor’s listing, privacy label, support documents and review feed, and never run by us. Describing someone else’s app with the confidence we describe our own is the fastest way to make these pages worthless.

First-hand covers the mechanics: how each of the seven categories detects what it detects and which thresholds you can move, what happens after you tap a tile, that a swipe marks a decision rather than deleting anything, that a keep is recorded and reversible from the Kept list, and that every deletion goes through the iOS confirmation sheet into Recently Deleted, where iOS holds it for 30 days. The per-category detail is on the feature pages.

It covers the free tier exactly, because we set it: nothing is locked, every category and every scan result is open to everyone, and only deletion is metered — 50 items of any type per calendar day plus 2 large videos, resetting at device-local midnight.

It covers the limits with the same precision. PhotoBroom has no album filing, no duplicate merge, no vault, and it never compresses or re-encodes anything — it only deletes. There is no Android build, and it needs iOS 18.0 or later. On an Apple silicon Mac and on Apple Vision Pro it is the iPad app running there, not a native build. Those sentences sit on comparison pages where they hand the point to somebody else, which is the only kind of first-hand account worth publishing.

It covers the network boundary, and we state that one precisely rather than roundly. Your photos and videos are analysed on your device and are never uploaded. There is no account and no sign-up. What we claim is that your media is processed locally — not that the app never talks to the network at all, which no app that sells anything can honestly say. The privacy page has the detail.

And it covers our own mistakes. PhotoBroom’s App Store description says the free tier is “50 photos and 2 videos a day”, which understates what actually ships — 50 items of any type, videos included. This site publishes the accurate rule. The listing is the thing that needs fixing, and saying so here is cheaper than letting you find the discrepancy yourself.

Screenshots on this site are PhotoBroom’s own App Store creative, so they show exactly what the listing shows. They are pictures of the interface and never measurements: any figures visible inside them belong to the library in the screenshot, and are not a forecast of yours.

Where the competitor facts come from

Every claim about a named app comes from one of five places: its US App Store listing, the App Privacy card on that listing, the vendor's own support or policy documents, Apple's public US customer review feed, or a dated third-party study. Where a source contradicts itself — and two vendors' own documents do — we publish the contradiction instead of picking the reading that suits us.

Unless a page says otherwise, everything was read on 3 August 2026. Prices, free-tier rules and privacy labels change without notice, so check the listing before you rely on any of it.

What we did not do

We did not install and benchmark these apps. Nothing on this site is presented as a measurement: there are no detection-accuracy percentages, no timing figures, no storage-reclaimed numbers. A benchmark we ran on our own competitors is not evidence you should trust, and inventing one would be worse than admitting we have none.

We also publish no competitor pricing. Prices drift, tiers get renamed, and a stale figure on thirty pages costs more credibility than a fresh one buys.

How user reviews are used

Where the only available evidence is the App Store review feed, the claim is written as a report — “users have reported” — and never as an established defect. Reviews are self-selected, skew negative, and describe builds that may since have been fixed. They are worth reading precisely because vendors do not publish their own failure modes, but they are not the same kind of fact as a privacy label.

Corrections

If something here is wrong about your app, tell us and we will fix it and say that we did. Get in touch.