A calculator laid out differently
The on-screen scientific calculator is not the one on your desk. Its key positions, its degree/radian toggle and its memory registers all have to be learned — and exam hall is a poor place to start.
In active development
Software for people
Software for people
sitting the hardest exams
A candidate spends a year on the syllabus and three hours on software they have never opened. We are building the rehearsal that removes that variable — the exact delivery interface, the real marking rules, and analytics that say what to fix next rather than just what went wrong.
- GATE EE & ECE first
- Designed to extend to other exams
- Works on a mid-range phone
Question 12
MCQ
+2.0
−0.66
A series RL circuit is excited by a unit step at t = 0. The current is governed by
L·di/dt + Ri = u(t)
The time constant of the circuit is:
(A) R / L
(B) L / R
(C) RL
(D) 1 / RL
Mark for Review & Next
Clear Response
Save & Next
Submit
A working mock of our test interface. The real engine adds the on-screen calculator, the numerical keypad and full keyboard lockdown.
The problem
Marks lost to the software are still marks lost
Ask anyone who has sat a computer-based competitive exam what surprised them, and almost none of it is the syllabus.
A keypad that refuses your keyboard
Numerical-answer questions accept input from an on-screen numpad only. Candidates who practised by typing lose time discovering that, and sometimes lose the answer with it.
A palette that means something
"Marked for review" does not save your answer in every state. Candidates learn that distinction the expensive way. In our engine they learn it in practice, where it costs nothing.
The engine
Fidelity first, content second
Most platforms build a question bank and wrap whatever interface is convenient around it. We build the interface to specification, then hang the content on it — because the interface is the part that transfers to exam day.
Countdown with auto-submit
The clock runs the way the real one runs, and the paper submits itself when it hits zero. Time pressure you have not rehearsed is time pressure you will feel for the first time on the day.
The on-screen scientific calculator
Same key layout, same degree and radian handling, same hyperbolic and logarithmic functions, same memory registers. Practised until reaching for it is automatic.
Numerical answers, properly
On-screen numpad only, physical keyboard disabled, and grading against a tolerance range rather than an exact string — so 0.333 and 0.33 are judged the way the board judges them.
Typeset mathematics and circuits
Equations, matrices and transfer functions render as mathematics, not as images of mathematics — sharp at any zoom, and legible when the question is read on a phone.
A dropped connection costs nothing
Responses are written locally as you go and re-synced when the network returns. An attempt should never end because a tower did.
The whole thing on a phone
A slide-over palette drawer, a thumb-reachable action dock and a numeric keypad that rises as a sheet. Not a desktop layout squeezed down — a layout designed for the device most candidates own.
Three question types, three sets of rules
Each behaves differently on input, on marking and on what a wrong answer costs. Practising them as if they were the same is how candidates lose marks they had already earned.
| Type | How you answer | How it is marked | Cost of a wrong answer |
|---|---|---|---|
| MCQ — single correct | One radio button from four options | Exact match against the key | Negative marking applies — typically a third or two-thirds of the marks |
| MSQ — multiple correct | Checkboxes; one or more may be right | All correct options and no incorrect ones | No negative marking, but partial selections usually score zero |
| NAT — numerical answer | On-screen numpad only; the physical keyboard is disabled | Against a tolerance range, not a single exact string | No negative marking — so never leave one blank |
Marking schemes vary by exam and by year. Ours are configured per paper from the official instructions rather than hard-coded, and the instruction sheet is shown before every mock — the same one you will be asked to accept on the day.
Analytics
A score tells you what happened. We want to tell you what to do.
Ninety-two per cent in Power Electronics and sixty-four in Machines is not a result, it is an instruction. The dashboard's job is to convert an attempt into a ranked list of things to fix, and then check whether fixing them worked.
Weak-area detection
An alert when your error rate on a topic crosses a threshold, with the practice set that addresses it attached.
Time-leak analysis
Where the clock actually went, split by question type — because two minutes lost on a question you answered correctly is still two minutes.
Question-level review
Every question, your answer, the worked solution, and the time you spent against the time it should take.
Progress against yourself
The only comparison that reliably means anything early on: this attempt against your last one, on the same subjects.
Content operations
A question bank is only as good as its audit trail
One wrong answer key in a mock test destroys a candidate's trust in every other question. So authoring, review and correction are a first-class part of the platform, not a spreadsheet somebody maintains on the side.
The author's desk
A formula editor that renders as you type, a tagging taxonomy down to sub-topic, difficulty grading, and source attribution to the year and paper a question came from.
The review desk
A second specialist audits each draft side by side with the original source, fixes formatting and tolerance ranges inline, and either publishes it or returns it with specific notes.
The correction desk
Candidates can flag a problem from inside a test without breaking their attempt. It lands on a live queue, gets resolved, and anyone affected is told what changed.
Worldwide
One engine, many exams
We are starting with India's GATE, in Electrical Engineering and in Electronics & Communication Engineering, because that is where our subject depth is and because its delivery software is exacting enough to be worth reproducing properly.
But the engine is not built around it. Strip any computer-based exam down and you find four things: a delivery interface, a question grammar, a marking scheme and a clock. Describe those four and the platform can rehearse a different exam, in a different subject, in a different country — without a rewrite.
We will not claim coverage we do not have. When an exam is genuinely supported, it will appear here by name with the date its specification was last verified. Until then this section describes a design goal, and we would rather say that plainly than list twenty exam badges nobody has tested.
Get involved
Three kinds of people we want to hear from
The platform is in build. That makes this the useful moment to talk, not the awkward one.
Candidates
Tell us which exam you are sitting and we will write to you once — when there is something real to try, not before. Early testers get the full platform free, permanently.
Join the listSubject specialists
If you can author and audit questions at exam standard — with the derivations intact — we want to work with you. Named authorship, and credit that stays attached to the question.
Talk to usInstitutions
Colleges and coaching institutes that want their own cohort running on a faithful interface, with results they can actually see. Pilot conversations are open now.
Discuss a pilotBe told when it opens — once, and only when it is real
No drip campaign, no launch countdown, no "join 10,000 aspirants" counter. One message, when there is something worth your time.