Screen time apps are usually compared on features: schedules, statistics, categories, widgets. That comparison is close to useless, because almost all of them do those things and none of it predicts whether the app is still running in a month.
The question that predicts it is narrower: at the moment you most want in, can you undo the block by yourself?
Four models, and what each survives
Friction — one sec, ScreenZen
A pause, a breath, an interruption screen before the app opens. You can always continue.
Survives: unconscious reaching, which is a large share of all phone pickups. Genuinely effective, and the least unpleasant thing on this list.
Does not survive: a decision you actually made. That is by design, not a defect — these apps are built on the belief that you should stay in control, and for many people that is the correct trade.
Timers — Forest, most focus apps
You start a session, and a countdown holds until it ends.
Survives: distraction inside a work block you chose to begin.
Does not survive: evenings and weekends, because you simply never start a session. An opt-in tool is only as reliable as the moment you opt in — and the moment you most need it is exactly the moment you will not.
Self-held passcode — Apple Screen Time, most strict modes
A real block, with a code you chose.
Survives: mild impulse. The prompt itself adds a useful beat.
Does not survive: genuine craving. You know the code. This is where the overwhelming majority of people fail, and it is a structural problem rather than a personal one.
Opal's Deep Focus belongs roughly here too — deliberately hard to reverse, and its own documentation suggests handing your phone to someone else to set a passcode you do not know. That works, and it is telling: the strongest version of a self-held lock is the one where it stops being self-held.
Another person holds it — Disconnect and a handful of others
The decision genuinely leaves your hands. Two versions exist.
The improvised one is asking a friend to type a passcode you never see. Free, works immediately, and collapses within a fortnight — not because the idea is wrong but because they have to be physically present every time you have a legitimate reason to get in.
The built-in one is a request-and-approve flow: you ask remotely, they answer in one tap. Same principle, without the logistics. Disconnect, LockPact, PeerPause and a few others work this way.
The comparison that matters
Laid against the single question:
- one sec, ScreenZen — yes, you can continue. Deliberately.
- Forest — yes, you can end the session.
- Apple Screen Time — yes, you hold the passcode.
- Opal — mostly no in Deep Focus, and stronger still if someone else sets your code.
- Disconnect (Hard lock) — no. Only your buddy can approve, and approvals expire on their own.
A fuller version of this, including free tiers and platforms, is on the comparison table.
None of these is the best app. They are answers to different questions, and picking the wrong one is why people conclude that blockers do not work.
The one you cannot cheat
Disconnect's Hard lock has no override and no backdoor — only your buddy can approve an unlock, and approvals expire on their own.
Get Disconnect →How to choose
Be honest about your actual failure mode, because they need different tools.
- You open apps without noticing. Friction. ScreenZen is free and very good at this. You do not need anything stronger, and stronger will feel like a punishment.
- You get pulled away during work. A timer, or a scheduled block during work hours.
- You lose whole evenings on purpose and regret it after. Nothing you can undo alone will hold. This is the case that needs another person, and it is the only case where it is worth the small social cost of asking.
Most people reach for the strictest tool for a problem that only needed friction, find it oppressive, and abandon the category entirely. Match the tool to the failure.