Why I Built My Own IPMAT Vocab App

This project was not supposed to exist.

I already have enough things I'm building.

EBO is still evolving.

SYNTHESIS is nowhere near what I eventually want it to become.

I have IPMAT preparation to do, which is arguably more important than either of those.

So naturally, while preparing for IPMAT, I started building another app.

But this one happened for a very different reason.

I wasn't sitting around trying to think of a startup idea.

I wasn't asking:

"What kind of education product could I build?"

I was just trying to study vocabulary.

And getting increasingly annoyed by it.

The vocabulary problem

The verbal section of IPMAT involves a ridiculous amount of language.

Vocabulary.

Synonyms.

Antonyms.

Idioms.

Usage.

Grammar.

Reading.

Words you recognise but can't properly define.

Words you've seen six times and somehow still forget.

Words that look almost identical but mean completely different things.

And then there are books with hundreds or thousands of them.

One of the resources I was using was Word Power Made Easy.

Another was BlackBook.

Both are useful in their own ways.

But eventually I found myself staring at the amount of material and thinking:

How am I actually supposed to retain all of this?

Reading a word once obviously isn't enough.

Highlighting it doesn't magically put it into my brain.

And going through enormous lists again and again doesn't tell me whether I can actually recognise the word when it appears inside a question.

That distinction started bothering me.

Knowing that I have "studied" a word and being able to use that knowledge during an exam are two completely different things.

Existing tools weren't exactly wrong

There are already vocabulary apps.

There are flashcard apps.

There is Anki.

There is Quizlet.

There are generic question banks.

There are websites with word lists.

I could have used any of them.

And technically, they work.

That wasn't the problem.

The problem was that I didn't want a generic vocabulary system.

I wanted something built around the way I was preparing for IPMAT.

I didn't just want:

Word → Definition

I wanted to be tested.

I wanted words to appear inside actual questions.

I wanted wrong answers to come back later.

I wanted to see what I was repeatedly getting wrong.

I wanted vocabulary, grammar and other verbal skills to eventually exist in the same practice system.

I wanted something I could open for a few minutes, solve a small set, learn something, and leave.

No setup ceremony.

No manually creating hundreds of flashcards.

No giant productivity system surrounding it.

Just practice.

Then I had the dangerous thought

At some point the obvious question appeared:

Why don't I just build it?

This is becoming a slightly recurring problem in my life.

Whenever software annoys me enough, my brain starts treating "make another piece of software" as a reasonable response.

A few years ago that sentence would have been ridiculous.

I didn't know how to build apps.

I still don't consider myself a programmer.

But the way I build things has changed a lot.

I can define what I want.

Break it into smaller problems.

Use AI to help me implement things I don't know how to code myself.

Test them.

Break them.

Change my mind.

Repeat.

So instead of spending another week searching for the perfect IPMAT vocabulary app, I started thinking about what my version would actually do.

I did not want to build Duolingo for IPMAT

The temptation when building an education app is to immediately add everything.

Levels.

Leaderboards.

Coins.

Avatars.

Daily quests.

Friends.

Chat.

AI tutors.

Achievements.

Ten different dashboards.

And suddenly the simple problem you were trying to solve disappears underneath a product roadmap.

I already learned that lesson while building EBO.

So this time I wanted the starting point to be brutally simple.

Open the app.

Choose what you want to practise.

Answer questions.

Understand what you got wrong.

Come back tomorrow.

That's it.

If the core loop wasn't useful, no amount of XP animations would save it.

The first idea wasn't even purely vocabulary

Something interesting happened almost immediately.

The more I thought about the actual exam, the less sense it made to create an app that only knew how to show vocabulary flashcards.

IPMAT verbal ability is broader than that.

Vocabulary matters, obviously.

But so do grammar, sentence usage, reading comprehension and other verbal skills.

If I was already building a question system, why artificially restrict it?

So the idea started shifting.

The personal problem that triggered the project was vocabulary.

But the product itself could become a broader IIM UG verbal practice app.

IPMAT would remain the reason I was building it.

It didn't have to remain the only thing the system could ever understand.

That distinction became important later.

I also didn't want to dump 20,000 words into a database

More content looks impressive.

If one app has 2,000 words and another has 20,000, the second one sounds better immediately.

But I started questioning whether that actually helps.

If I'm preparing for a specific family of exams, I don't necessarily need every obscure English word someone managed to collect.

I need relevant vocabulary.

I need repeated exposure.

I need context.

I need recall.

And I need to know which words I'm weak at.

A massive database can make the app look comprehensive while making the actual learning experience worse.

So before building a giant word library, I wanted to understand what deserved to be inside it.

That decision eventually led to much more research than I expected.

But at the beginning, the goal was still very small.

Build something I would actually use

This became the main rule.

Not:

"Would this look impressive on my portfolio?"

Not:

"Could this become a startup?"

Not:

"How many users could this have?"

Just:

Would I genuinely use this while preparing for IPMAT?

That question is surprisingly useful.

Because I am the exact person the first version is for.

If opening the app feels annoying, I know immediately.

If a practice session is too long, I feel it.

If an explanation is useless, I notice.

If a feature sounds cool but doesn't help me study, I can remove it.

There is no imaginary target user yet.

I'm sitting right here.

The first version could be terrible

I was completely fine with that.

I didn't need accounts.

I didn't need perfect cloud sync.

I didn't need Android, Windows and iOS apps.

I didn't need thousands of questions.

I didn't need adaptive AI.

I didn't need a beautiful landing page.

The first version only needed to answer one question:

Is practising this way actually better for me than continuing with the system I already have?

So I decided to build the smallest version capable of answering it.

A tiny question bank.

Short practice sessions.

A way to see mistakes.

A way to return to weak questions.

Enough structure that I could start using it myself.

And nothing more until that worked.

At least that was the plan.

Because, as usually happens with these projects, the tiny app did not stay tiny for very long.

But before it became what it is now, I had to figure out something much more fundamental:

What should I actually be learning in the first place?