Self-Discipline App: What One Can and Can't Do for You
A self-discipline app can't make you do anything, with one exception: it can refuse to open something on the phone it runs on. So sort your problem first. If you want to stop something on a screen, such as a feed, a game or a site, you need a blocker. If you want to keep something going off the screen, such as practice, study or training, there's nothing to block. For that, an app can do three plain jobs: hold one small daily commitment you set in advance, ask you every night whether you kept it, and give you a way back after a miss that leaves the miss on the record.
This site belongs to Oathcraft, an app of the second kind and not the first. It holds one daily oath, asks you to close each day with a true answer, offers a repair after a broken day, and blocks nothing. It's free to use, with ads. You can hold any app to those three jobs, this one included.
What should you look for in a self-discipline app?
Take one thing you'd like to keep going. Say it's the piano. You want to play again, and you've settled on a small version: once the kids' light is off, you put the headphones on and play one piece through. A blocker has nothing to hold here. Nobody has written the app that locks you out of the couch.
A commitment like that usually dies at one of three ordinary moments. An app can do one job at each, and you can check for each job before you rely on it. A store listing won't show you any of the three. The first week won't either, because that week you'd likely have played anyway.
When it comes due. The light goes off, you're tired, and playing one piece through becomes looking for a new piece to learn, which you can do from the couch with your phone. You never decided to skip. The terms moved. The first job is to hold the terms you set when you weren't tired: what you'll do, worded so it's plainly done or not done, and tied to a moment in your day so you aren't picking a time each night. To check an app, set up one commitment, then try to change what today requires, today. If it lets you, it's only keeping your latest offer.
When the day ends. A skipped day mostly goes unrecorded. Nobody asks, the day ends without comment, and by Sunday the week has become "I played a few times". The second job is to ask every night, by a deadline, and to have a place for no. A missed day should go in as an entry the way a kept one does, and a day you don't answer for should count as something too. Check what the app does with a blank. Then check whether you can go back and tick last Tuesday. If you can, the record will drift toward the week you wish you'd had.
There's evidence for the asking. In 2016 Benjamin Harkin and colleagues pooled 138 experiments, with 19,951 people between them, in which some were prompted to keep track of their progress toward a goal and others weren't. The groups prompted to track got further toward their goals on average, and the effect was larger when the progress was physically recorded. That's an average across all 138 experiments and not a test of any one app, so take it as support for the asking, not for a product.
The day after a miss. You didn't play on Tuesday. On Wednesday a tracker often leaves you two ways to go on: mark Tuesday as if it went fine, or start again from zero. The first is a small lie. The second makes one tired night cost every night before it. The third job is a way back that's neither: the miss stays in view, there's one specific thing to do about it, and there's a time limit for doing it. To check, find out what the app shows a week after a miss. If it has a freeze or a pause that keeps a count going, ask whether you can apply it once the day is over, and whether the day still shows as not done. A cover you can add afterward is the tick on last Tuesday under another name. If the only way on is a counter at zero, ask yourself honestly whether you keep going from zero. If a zero has ended things for you before, the guide to a broken streak has more on the next step.
How does Oathcraft do those three jobs?
Put the piano into Oathcraft. On the first day there's no account to create. You take the First Oath, three lines that open with "I will honor my word" and end with "I will return, as many times as it takes." Then you write a single daily action, "play one piece through", and tie it to a cue, either a routine or a time. Here that's the kids' bedtime. The app shows it back as a sentence that starts "Every day, before I end my day". The cue is yours to notice; the app doesn't send a notification for it. You can edit an oath later, but the edit takes effect from the next practice day, so tonight's terms stay as written. That's the first job.
The second is the Close. If you allow notifications, one reminder asks you to close the day, at 9 p.m. unless you move it. On Monday you played, so you mark the oath done and Monday goes in as Full. The app can't check that. It takes your word. It can ask, and the true answer has to come from you. Your first Close needs no account. To close the next day you need a free one, with an email and a password.
On Tuesday you didn't play. You have until 3 a.m. to close the day, and it can end in one of three ways.
You never open the app. At the cutoff Tuesday becomes a Break by default, and Run and Streak, the two counts the app keeps, go to zero. You can record yesterday during the following day, but only if you'd done the oath before the cutoff, and you hadn't. After 3 a.m. a day can't be changed, so there's no ticking Tuesday on Thursday.
You close Tuesday yourself and leave the oath unmarked. That's a Break too. A new user has nothing between Full and Break, so half a piece would also be one. But a Break you've saved can be repaired. You choose Try again, or you make the oath smaller or clearer, say the first page of the piece. Save the repair before the cutoff that ends Wednesday and both counts hold at Monday's numbers. Tuesday adds nothing to them. History goes on listing it as a Break, with the repair next to it. The hold works for a single Break, the first since a successful day. If Wednesday is a Break as well, both counts reset even with a repair.
Or you saw Tuesday coming. One day in any seven, you can close the day as Rest: a pause you declare that same day, before the cutoff, with nothing marked done. Rest holds Run and Streak where they were and adds nothing, and the day goes into the record as Rest. On Wednesday it's too late to make Tuesday a Rest. And after a Break, a Rest doesn't count as your return: a second Break before a successful day still resets both counts.
Nothing else holds a count. There are no freezes or tokens, and nothing you can pay for to rescue one.
One thing changes later, and it matters for the piano. Seven successful days in one Run bring you to Rank I, and they don't have to be consecutive, since the Run holds through a Rest and through a Break repaired in time. At Rank I the app has you set up a Core: a small daily commitment for sleep and recovery, which sits next to your original oath. From then on the Core decides the day. Both kept is Full. The Core kept and no piano is recorded as Core, which still counts as a successful day. Only a missed Core makes a Break. Until then, the practice is one oath and one question a night.
What doesn't Oathcraft do?
It doesn't block anything. You could make "the phone stays in the kitchen overnight" your oath, and the app would ask about it every night, but nothing would stop you walking to the kitchen.
It has no other people in it: no partners, friends, sharing or leaderboards. There's nobody to report to.
Every oath is daily. There's no setting for weekdays or for three days a week, and the only day off is that one Rest in seven. A commitment with more days off than that built in needs a different kind of tracker; Habit Tracker: How to Choose One You'll Keep Using sorts those by what they record. Extra daily oaths are allowed, but before Rank I a single missed one turns the whole day into a Break.
It doesn't spare you every zero. A day you never close, a Break you don't repair in time, or two Breaks running will put Run and Streak at zero. Rank, once earned, doesn't go down.
It runs on iPhone and Android, and there's no version you use in a browser.
So it suits someone with one off-screen thing to keep going every day, who wants to be asked and is prepared to answer truthfully. If your trouble is a screen, get a blocker first. If you'd only answer truthfully with someone else reading, you want an app with people in it, or a person.
Is a self-control app the same as a self-discipline app?
Usually not. In the app stores, "self-control app" usually means a blocker, and that's the right tool when the evening you want is one where you don't open the video app in bed. Oathcraft is no use there. Look for a blocker, or begin with the screen-time limits already in your phone's settings. What matters most is how hard the block is to undo in the moment. A limit you can lift with one tap is a reminder with an extra step.
Some stops live off the screen, like staying up too late or buying lunch out every day. Nothing can block those. They have to be reworded as something you do, such as lunch packed before bed, and then they belong with the piano. Plenty of people have one of each kind, and then two tools make sense.
What if you only follow through when someone is watching?
Then you need a person more than an app. The pooled experiments above point the same way: keeping track helped more when the results were reported to someone or made public.
The plain version is one friend and a text each night that says yes or no. Some apps build this in, with partners or groups, and a few add money you lose when you miss. If an audience is what moves you, look for that in a listing. Oathcraft won't do it, because there's nobody in it to tell. Nothing stops you using both: a private record you answer to, and one person who gets the same answer.
Is there a free self-discipline app?
Yes. A good many are free to install, and the store listing says whether there are in-app purchases. What's worth checking is which of the three jobs the free part covers. Being asked every night for free and paying for charts is one bargain. Paying for the way back after a miss is another.
In Oathcraft, oaths, the Close, Repair, Rest, History and the ranks you earn are free at every rank, with no end date. The free version shows ads, though not video or full-screen ones. An optional paid upgrade, Oathcraft Plus, removes the ads and adds Counselor, which is AI guidance for obstacles and next steps, along with guided preparation, Close reflections and weekly reviews. The App Store or Google Play shows current prices.
The cheapest version is paper. An index card on the music stand can hold the sentence, and a wall calendar where you write yes or no before bed can take the answer. For the day after a miss, write what you'll do about it in the next square. Paper sits where the thing happens and has nothing else on it. What it can't do is ask. The question only arrives if you remember it, and the deadline is yours to move.
Can an app give you self-discipline?
No. Besides blockers and trackers, the stores file two more kinds of app under the word: lessons with daily quotes, and coaches that message you. Both work with words, so what matters is when the words arrive.
A lesson or a quote acts while you're reading it, and reading about discipline can feel a lot like having some. The moment you were worried about comes later, in another room, with the phone face down. A coach's message can land at that moment, the way a reminder can, and it can be swiped away the same way. Lessons and coaching earn their place when you don't yet know what to do: how to practice, which piece is a sensible first one. If you know exactly what to do and it's the doing that fails, more instruction isn't the missing part. That part is the subject of How to Build Self-Discipline Without More Willpower.
No app of any kind can want the thing for you on a night when you don't. What's left for one to do sounds like office work: keep a sentence, ask a question, file the answer. If you've deleted a few of these apps and decided the fault was yours, run the three checks on the last one before you settle on that. It may never have asked.