Choose one of your games
Enter your Chess.com or Lichess username and choose a recent public game.
Build a repeatable decision process you can use in real games.
All you need is your Chess.com or Lichess username.
Understand why you made the mistake—and learn a better way to think through the next decision.
Enter your Chess.com or Lichess username and choose a recent public game.
Solpento analyzes your moves and opens one serious, teachable mistake.
Work through your reasoning and finish with one concrete thinking habit to practice in your next games.
Beyond the evaluation bar
The engine gives you the answer. Solpento helps you improve the thinking process that leads to it.
“18. Bb5? −2.1. Best was 18. Qd2.”
“Your plan made sense. You considered what your move would achieve, but not your opponent’s strongest reply. Next time, test your idea before committing.”
Solpento uses mistakes from your own games to help you build a repeatable process for evaluating ideas and deciding what to do next.

No. A good human coach can understand your broader goals, follow your progress, and adapt training over time. Solpento is designed to complement that work—or help when regular coaching isn't available—by turning one important mistake from your game into a focused guided review.
Solpento uses an LLM to guide the conversation, but it does not rely on the model alone for chess facts. Positions and moves are grounded in Stockfish and structured chess analysis, while the conversation is carefully guided around the lesson. Mistakes can still happen, especially during beta, and we're continuously improving the safeguards and review quality.
No. Solpento starts with a mistake from your own game—not a position chosen because it contains a puzzle. You begin by explaining your idea before receiving progressive guidance.
Enter your Chess.com or Lichess username and choose from your recent public games. You don't need to connect your account or share a password.
Solpento is designed primarily for online players around 600–1400 who know the rules but want a more reliable process for making decisions in real games.
Review the decision, not just the move