You do not need a new app to find out where your evening went. Android has been keeping that record the whole time, and it is sitting in Settings, two taps in, on almost every phone sold in the last six years.
The reason people ask anyway is that the numbers are buried and the labels vary by manufacturer. So here is where the data actually lives, what it counts, and what to do once you have read it.
The built-in dashboard
On stock Android and Pixel:
- Settings → Digital Wellbeing and parental controls → Dashboard.
- The bar chart shows time in each app for today. Tap a bar for an hour-by-hour breakdown.
- Tap the chart’s title to switch between today, yesterday, this week and last week.
That is the whole trick. It is the same data a paid screen-time app would show you, because it comes from the same place.
Manufacturers rearrange this. Samsung ships Digital Wellbeing as a separate app in the launcher. Some ROMs put the equivalent under Settings → Battery, or inside a vendor “phone manager” or “security” tool. If you cannot find it, search Settings for the word “wellbeing”, then “usage”, then “screen time” — one of the three will land you on it or on something that links to it.
The permission underneath it
Every one of these screens is reading the same system service: UsageStatsManager. It records when apps move to the foreground and when they leave, and it keeps those events for a while.
An app can only read it if you grant Usage access, which on stock Android is at Settings → Apps → Special app access → Usage access. This is a genuinely sensitive permission: it is a complete log of which apps you open and for how long. That is exactly why Android makes it a separate, explicit grant rather than part of install.
Worth knowing: granting Usage access to an app is not the same as that app storing anything. The permission governs reading. What happens next is the app’s policy, and it is the first thing worth checking in any screen-time app’s privacy page.
What the numbers do and do not count
Once you have the dashboard open, the figures are more specific than they look.
- Foreground time, not screen-on time. Ten apps open for one minute each is ten minutes of app usage. If the screen was on but showing the home screen, that is generally not attributed to an app.
- No launcher, no lock screen. Time spent deciding what to open is invisible, and for some people that is a meaningful share of the evening.
- Sessions are cut into daily buckets. A session that crosses midnight gets split, which can make a late-night scroll look smaller on both days than it felt.
- Unlocks and notifications are separate series. A high unlock count with low app time is a different problem from a low unlock count with two hours in one app. The first is habit, the second is a single app that has your attention.
That last distinction is the useful one, and it is why “screen time” as a single number is close to meaningless. Three hours split across a dozen deliberate uses is not the same three hours as one continuous feed.
Why per-app time beats a daily total
A daily total is a verdict. Per-app time is a diagnosis.
If the total says four hours, the only available response is guilt, and guilt is a poor design input. If the breakdown says forty minutes in a messaging app you actually needed and two hours in one short-video feed, you have a specific, small, solvable problem. Android already tells you which app just came to the foreground — that signal is what makes per-app measurement possible in the first place.
This is also why hard limits fail so often. “No more than three hours of phone” fights every legitimate use at once. “Not this app, not right now” is the version that survives contact with a real evening.
What to do with the numbers
Three approaches, in increasing order of effort:
App timers. Digital Wellbeing includes per-app timers, and when the time is up the app is paused for the rest of the day. The problem is that the pause is a dialog, and a dialog is a thing you can dismiss.
Scheduled quiet. Some phones and most launchers can turn off notifications or greyscale the screen for a window. This works well for bedtime and badly for the middle of a workday.
Friction at the moment of opening. Instead of blocking an app, insert something small between the tap and the feed. A five-second pause is enough to break the reflex, because the reflex is what you are fighting — not the app, and not your own willpower.
That third approach is what ScrollBreak is built around. You choose the apps that pull you in, and after the interval you set — one, two, five, ten, fifteen or twenty minutes — the app is interrupted with a short challenge: a jigsaw puzzle, a coloring template or a breathing exercise. Solve it and you are back in. Nothing is deleted and nothing is permanent. Difficulty steps up gently, so the break stays a speed bump rather than becoming a second job.
Before you install anything, the roundup of screen time apps for Android compares the options for each kind of habit, including the Digital Wellbeing that is already on the phone.
The measurement side is deliberately boring. Foreground timings come from UsageStatsManager on the phone, the accessibility service is used only to notice when a protected app becomes active and when it closes, and usage history stays in local storage. No account, no cloud sync. Which is the only version of “private” worth claiming.
Being straight about the trade-offs: the free plan protects two apps and shows a banner on the Challenges tab plus an optional rewarded video. Pro removes the ads, lifts the app limit and adds schedule windows. If two protected apps is enough for you, the free version is genuinely usable and you never have to watch an ad to use the thing you installed.
The one number to look at first
If you only open the dashboard once, sort by time and look at the top entry. Not the total, not the pickups — the single app at the top of the list.
For most people it is not the app they would have guessed, and it is almost always the one they would not defend out loud. That app is the whole conversation. Everything else on the screen is context.