Two years ago, we were those parents.
Driving to tournaments every weekend. Paying for the classes. Watching our kids play hundreds of games online. And having no real idea if any of it was working.
The coach would say "practice tactics." Chess.com would serve up random puzzles. Nobody could tell us the one thing we actually wanted to know: what, specifically, does my kid struggle with — and what should they do about it?
So I built it.
CoachMyChess takes a kid's real games — imported from Chess.com and Lichess, or even a photo of a board — and turns them into feedback the kid can actually understand and act on. It finds the patterns in how they lose, builds practice sets around their real weaknesses, and gives parents a simple dashboard. No chess expertise required, because I don't have much either.
A few things I believe more strongly after building it:
Communication is the hard part, not the AI
Getting a model to spot a recurring blunder is the easy 20%. Explaining it to a 7-year-old in a way that's encouraging instead of discouraging — that's the 80%. If a kid feels judged, they close the laptop, and then none of the cleverness matters.
Build for the person standing next to the user
In chess, the player is the kid. But the buyer, the worrier, and the weekend driver is the parent. A tool that ignores the parent doesn't get used, no matter how good it is for the kid.
Your harshest users are a gift
My 6th grader and my 3rd grader are my QA team. They find every rough edge and they are not polite about it. If a feature survives them, it's ready for yours.
I didn't build CoachMyChess to disrupt anything. I built it because we needed it and it didn't exist. That's still my favorite reason to build.