For students
See the map. Trust the next tap.
Home shows strength, coverage, and one next action with the same why the planner stored. You are not guessing which chapter to grind.
JEE Main · JEE Advanced · AP & TG EAPCET
You do not need more questions. You need one highest-leverage next action — and a reason that updates as you work. Klarity names the trap, the guess, the stale topic, or the weak prerequisite. The tutor only explains.
TypeScript picks the move. Parents see counts and tones — never the chat.
For students
Home shows strength, coverage, and one next action with the same why the planner stored. You are not guessing which chapter to grind.
For parents
Under 18 links a parent login (DPDP). Ten seconds: time, week score, latest check, and what to ask about next — never stems, hints, or tutor chat.
EAPCET · JEE Main · IIT
The goal is a rank on a specific paper — not another chapter playlist. Klarity spends practice on the leak that costs marks on that paper. It does not promise a percentile or predict tomorrow’s question.
AP & TG EAPCET
EAPCET rewards coverage and clean speed. If vectors still leak, projectile items waste the sitting. Klarity fixes the prerequisite first, then drills guesses instead of handing a harder set.
Paper-style sets stay on their own score. A topic check is understanding — not your EAPCET standing.
JEE Main
Negative marking punishes the same trap twice. A named misconception becomes REMEDIATE, not another random cousin. Slow but correct becomes timed practice — Main is a clock.
A 6-second lucky hit does not raise Main readiness. Only clean sits do.
IIT · JEE Advanced
IIT is JEE Advanced: multi-concept items, traps, and pressure. Klarity will not open that set until the chain is ready. Stale “strong” from last month comes back as review before a harder archetype.
Exam readiness is a separate meter from topic score. The UI never invents an IIT percentile.
Student home · Parent glance
The student gets a skill map and a visible why. The parent gets counts and a tone. Isolation is in the API — not a hidden button.
Student
Home and the topic hub show the same stored reason. Strength and coverage update after clean attempts. A topic check is understanding. Exam readiness is a separate meter — and stays quiet until a clean sitting exists.
Parent
Time on the app, week correct / items, latest check, and a glance — on track, needs a look, or a quiet week. If they rated “how sure?”, you see over / under / calibrated. If they sat a paper, readiness appears. Help is a count, never the wording.
How it works
Not one subject percentage. A short test on a topic and its prerequisites — so every concept gets a live label: untouched, weak, developing, strong, or mastered.
A tab leave, a hint, or a 30-minute idle sit does not count as fluency. Only clean attempts feed time bands, guesses, and readiness.
A wrong choice can write a trap code. Fast misses look like guesses. A strong label with an old last-correct is due for review. Weak vectors block projectile motion.
The planner stores the action and the reason. Home and the topic hub show that string. The LLM tutors the item — it does not pick TEACH vs REMEDIATE vs REVIEW.
Intelligence
Competitive students fail the wrong next ten minutes. Klarity spends that time on the move that changes standing.
Range formula on a flight-time stem writes a code. Next action is repair that trap — not another random cousin.
Fast and wrong twice is an accuracy drill. Slow and correct is timed practice. A 6-second lucky hit does not raise exam readiness.
Last-correct is not last write. A strong label can still enqueue review. The stored label stays; only the next move changes.
92% on the topic check is understanding. JEE-style sittings write a separate readiness meter. The UI never invents a percentile.
When we ask “how sure?”, that rating sits next to correct or miss. Parents see over / under / calibrated — we do not invent confidence from the clock.
Minutes, week score, latest check, what to ask about next, calibration if they rated certainty, readiness if a paper exists. Isolation is in the API — not just a hidden button.
Coverage
One coach for Physics, Maths, and Chemistry — JEE Main, JEE Advanced (IIT), and AP / TG EAPCET — so you are not stitching three apps together. Target exams on your profile steer review, not a predicted question.
What you get
Per-concept labels plus the planner reason. Home, Plan, and the topic hub show the same next action — so you know why this tap, not another chapter.
Rate how sure you are on some items. Overconfidence is a tone from those ratings — never inferred from how fast you clicked.
Mastery, fluency, and next action are TypeScript rules. The LLM explains and hints. It never writes the skill map or the why.
Under 18 adds a linked parent (DPDP). They see time, week score, latest check, next topic to ask about, calibration, and readiness if a paper exists — never hint text or tutor chat.
Why this is not another coach app
| Typical app | Klarity |
|---|---|
| Another random question and call it personalization. | One highest-leverage next, with the same why on home and the topic hub. |
| A 6-second correct looks like mastery — or exam ready. | Dirty clocks and sub-10s hits never move fluency or readiness. |
| Right / wrong and a chapter percentage. | Five labels, a named trap, guess vs slow-correct, and a due review. |
| A chatbot that answers whatever you type. | A Socratic ladder on the current item. The model does not pick the next move. |
| Topic score treated as JEE readiness. | Topic understanding and exam readiness stay on different meters. |
| Parents see chats or nothing. | Parents see counts and tones. Isolation is in the API, not just the UI. |
FAQ
One highest-leverage next — and a why that updates. Physics, Maths, and Chemistry.