Kasturi

A few month back, I noticed a gap while validating my idea. AI would give different answers every time, competitors were hard to map, and public data was scattered across dozens of platforms.

So instead of relying on prompts, I built something for myself a small tool that combined AI reasoning with real signals from the market.

That gradually turned into MarketScope.

What I tried to do-

The first version of MarketScope was fast and useful it compared new startup ideas with existing products using public data, demand patterns, and market gaps.

It was meant to be a quick scanner to help founders avoid duplicate ideas and save hours of research.

But there was a problem:

AI often gave optimistic scores because it didn t consider real-world execution issues.