Chess Coach AI Is Live. Now I'm Not Sure What It Wants To Be.
Chess Coach AI is live and actually works, but I'm still figuring out whether this is a real brand I should build around or just a product I wanted to ship.
Chess Coach AI Is Live. Now I'm Not Sure What It Wants To Be.
Chess Coach AI is live at chess.benbravo.net, which is cool mostly because it means this thing is finally real.
It has an interactive board, in-browser Stockfish 18 analysis, opening explorer data, AI chat that can actually interact with the board, PGN export, annotations, multiple chat tabs, Google sign-in, premium, and rating-aware coaching. So this is not just some hacked together demo anymore. It works.
The main problem now is the AI itself.
My rough estimate is that the coaching starts to get shaky around 1500. Not unusable, just flatter. In simpler or tactical positions, it can still be helpful. But once the chess gets more strategic, the explanations start getting generic right when they need to get sharper. That is the part I still do not know how to solve cleanly.
And that leaves me in a weird spot.
Because on one hand, I am actually proud of the product. It does a lot already, and getting all these pieces to work together was not trivial. On the other hand, I still do not know what this is supposed to become. I cannot tell if this is something I should really commit to with a logo, branding, and a full polished identity, or if it is better as a product I keep building because I think it is interesting.
Right now it feels like both.
It is live, it is useful, and I think there is something here. I am just not sure yet whether Chess Coach AI is becoming a real brand, or whether it is still just me pushing on a weird technical problem I care about.