Chess Coach

• 

  icon

Inspiration

My 7-year-old tears through chess games without stopping to think. Every loss taught him nothing because no one was there to ask why. That gap is what Chess Coach was built to close.

What It Does

Chess Coach sits beside the child as they play — evaluating every move with Stockfish, then having the AI explain what was good, what was missed, and why. The child speaks their reasoning aloud and the coach pushes back. At the end of the session, Nutrient DWS builds a PDF report — move by move, scored and annotated — that a parent signs and keeps.

How We Built It

Flask backend, Stockfish at depth 9 for move evaluation, and an LLM for coaching commentary and voice dialogue. Nutrient DWS converts the session data to a structured PDF at the end of every game.

Challenges We Ran Into

Each move triggers multiple API calls in sequence — Stockfish, the LLM, TTS — and the latency stacks up. Keeping that pipeline fast enough that a child doesn't lose interest was the hardest engineering problem.

Accomplishments We're Proud Of

A 7-year-old sat through a 68-move game, spoke his reasoning aloud on most moves, and asked to play again. The PDF that came out of that session is in this submission.

What We Learned

Kids don't need to teach everything, just show them the way and spark interest, they will learn themselves

What's Next for Chess Coach

Take it to chess clubs and after-school programmes. The PDF report gives parents and coaches something tangible — that's the wedge into an institutional market that the big players have largely ignored

Built With

• openai
• python