Home

LockedInLearning Wiki

LockedInLearning is a home-screen replacement (launcher) that makes you answer a study question correctly before it lets you back into your apps. Instead of a passive habit tracker, it puts a hard checkpoint between you and your phone.

Pages

• CSV Import Format — how to format a CSV to bulk-import questions into a deck.
• Failure Modes — what happens when you answer wrong, and how to configure it per deck.

Why use it

Most study/habit apps are opt-in: you have to remember to open them, and it's just as easy to not. LockedInLearning flips that — the friction is on the path back to distraction, not on the path to studying. Every home-screen return becomes a tiny, unavoidable review rep, which is a much stronger forcing function for spaced repetition than a notification you can swipe away.

Best practices

• Start with a small, high-quality deck. A deck of 20-50 well-written questions you actually want to drill (vocab, formulas, facts) works far better than dumping an entire textbook in — every gate interruption should feel worth answering.
• Match the failure mode to the deck's content, not just to how strict you want to feel — see Failure Modes for the full breakdown:
  • MCQ and short factual answers — safe with any mode, including HARD_LOCK, since there's always a definite right answer to converge on.
  • ⚠️ Do not use RETRY (unlimited attempts) with open-ended question banks. RETRY never bypasses — if a FLASHCARD/free-text question's grading doesn't line up with how you naturally phrase the answer (it only tolerates small typos via a Levenshtein distance of ≤2, not rephrasing or partial credit), you can get stuck retrying an unanswerable question forever with no way through. Use MAX_ATTEMPTS (which lets you bypass after a few tries) or TIME_PENALTY for open-ended/flashcard-style decks instead.
  • Reserve HARD_LOCK for decks where you're confident every question is unambiguous — it has no bypass path at all once attempts run out.
• Keep hints on for decks in early "learning" stages, and consider turning them off only once a deck is mostly "mastered" (see mastery levels in Failure Modes).
• Set a sensible cooldown. The gate won't reappear for a while after a correct answer — tune this so you're not answering the same question every 10 seconds, but still often enough to matter.

Limitations

• Backgrounded apps can bypass the gate. The gate only intercepts you when you return to the home screen. If you already have an app open and switch to it via Android's Recents/Overview screen (rather than going home first), you go straight back into that app — the launcher doesn't (and can't, without accessibility-service-level hooks) intercept task switching directly.
• This is a launcher, not a device-wide blocker. It governs what happens on your home screen; it isn't a parental-control or app-blocking tool, and a determined user (i.e., yourself) can always switch the default launcher back in system settings.
• Grading has no semantic understanding — it's case-insensitive exact matching, with light typo tolerance only for flashcards. MCQ and LANGUAGE questions require a case-insensitive (and for LANGUAGE, accent-insensitive) exact match to the stored answer; MATH requires an exact numeric match; only FLASHCARD free-text answers get any fuzziness (Levenshtein distance ≤2, so small typos pass but rephrasing or partial answers don't).