LairsFlow trains real exam pacing with a per-question timer, turns every attempt into structured data, and hands it to an AI advisor that's grounded in your actual numbers - not guesses.
Students lose marks not from lack of knowledge but from lack of time-awareness - burning eight minutes on a two-mark question, then rushing the ones they knew cold. A Pomodoro timer has no opinion on any of that.
Before every question you set a time target - say, 3 minutes. LairsFlow logs both the target and the real time taken, then folds that gap into every insight it gives you, from the dashboard to the Advisor chat.
Pick a chapter, dial in a target for this one question - minutes:seconds, not a generic block.
Solved early? Tap to stop. Time's up? That's not a failure - it's just data on your real pace.
Difficulty, resource used, mistake type - conceptual, calculation, silly, or time management.
Save & Next opens the next timer immediately. No interruption, no context-switch.
Chapter selection. Chapters are subject + topic pairs - e.g. Physics → Electrostatics. Add them one at a time or bulk-import a whole syllabus via JSON, so setup for a new exam attempt takes minutes, not an evening.
The timer screen. A large countdown is the centerpiece. You edit the per-question budget in minutes:seconds before you start. While a session is running, a floating "Dynamic Island" stays visible with the current chapter, total session time, and current question time, plus a one-tap stop.
Solved early vs. time's up. Tap the timer to stop it the moment you finish - that opens the log screen with "solved within budget" pre-filled. If the timer hits zero first, nothing bad happens: it just means this question took longer than your target, and you log the real outcome.
The log screen. The core fork is "Did you solve it?" From there you record difficulty (Easy / Medium / Hard), the resource you used (textbook, coaching notes, video, custom), and - if you didn't solve it - the outcome (wrong answer, didn't attempt, conceptual error) and mistake type (conceptual, calculation, silly error, time management). Quick-note shortcuts keep this fast.
Undo. Mis-tap or mis-log something? The undo button reverses the last entry without breaking your flow.
Every field you fill here - timerSetting (your target) and timeOnQuestion (what actually happened) - becomes a row in your question log, which is exactly what powers the dashboard's charts and every Advisor answer.
A dynamic island tracks chapter, session time, and question time while you work - with a one-tap undo if you mis-log.
Every attempt records outcome, difficulty, resource, and mistake type, so patterns show up instead of vague feelings.
Activity heatmap, day-wise accuracy trend, and a question-quality breakdown - filterable by subject, chapter, and date.
Set a target date once and every stat - including the Advisor's - starts weighing days remaining against chapters left.
Ask for a nudge and it's scheduled as a real notification - the Advisor won't drop it inside a stated busy window.
Sign in once and your log, chapters, and Advisor conversations follow you across every device.
Every logged question lands here automatically - no separate tracking, no spreadsheet. Filter by subject or date range and the same numbers power your Advisor conversations.
Every reply is computed from a fresh snapshot of your question log, accuracy, and pacing - the same math the dashboard uses. Nothing is guessed.
If you're signed in, chat history and Daily Reports - including anything you type about distractions or how you felt - are stored in Firestore, not just kept on-device. A banner in the chat panel says this explicitly, every time.
The Context menu toggles themselves are a per-device preference and are not synced across devices - you set them fresh wherever you're studying.
Clearing a chat (trash icon) only wipes local chat history; your saved targets and Daily Reports in the Reports tab are untouched.
LairsFlow Advisor
grounded in your data
Tap "Set today's target" and it weighs stale chapters, weak subjects, and your pacing goal into one concrete plan.
It already knows today's real completion numbers, so it only asks 2–4 short questions about what it can't know.
Ask specific questions and get answers built on your real computed numbers, not the model's guesses.
With a pacing goal set, every answer factors "days remaining vs. chapters remaining" automatically.
Describe a plan and it schedules a real notification - in-app, and to your phone if paired - reasoning about timing.
With Phone Companion toggles on, advice factors in real screen-time and distraction-app data, not just self-reports.
Pair an Android device from the bell icon by entering the Sync ID shown in the companion app. Once paired, the Advisor always knows whether your phone is synced - even with every toggle below switched off - but it won't see any actual content until you turn a category on.
This is what lets the Advisor move from coaching on what you typed in a Daily Report to coaching on what your phone actually recorded - real screen time instead of a self-reported guess.
Of the four Phone Companion categories, recent phone notifications is the most sensitive - it includes message previews from apps like WhatsApp. We'd suggest turning on screentime and app-usage data first, and only adding notifications if you specifically need the Advisor to reason about a distraction that showed up as a notification, not just an app open.
Every category here is independent - enabling one doesn't enable the others, and you can turn any of them off at any time from the Context menu.
Yes. The Advisor doesn't ship with a shared backend key - you get a free one from Google AI Studio and paste it into Advisor settings → Advisor Configuration. You can save multiple keys and switch between them if you hit a free-tier limit on one.
10 a day, tracked per device. Only prompts that get a successful reply count against the limit - a failed call or a Retry that fails doesn't burn one.
Only what's toggled on in the Context menu above the chat input: Profile, Question Log, Reports & Pacing, and App Knowledge are on by default. Raw Data and Phone Companion start off since they're either heavy on tokens or sensitive. If it genuinely needs something that's off, it asks - you enable it inline and it re-answers immediately.
No - it's entirely optional and every toggle starts off. Pairing only tells the Advisor whether a phone is synced; it needs a specific toggle on before it sees any screentime, app-log, device-status, or notification content.
Yes - chapters are just subject + topic pairs you define, so the same per-question timer and logging loop works for any syllabus-based exam, competitive or otherwise.
Nothing - "Clear chat" only wipes local chat history for that conversation. Your saved targets and logs in the Reports tab are untouched, and the action is device-local, so it doesn't affect sync elsewhere.
If you're signed in, your question log, Daily Reports, and Advisor chat history sync via Firestore across devices. If you're not signed in, everything stays local to that device. Context menu toggles are always device-local, signed in or not.
Free to start. Bring your own Gemini key when you're ready for the Advisor.
Start free