Work status

The current product status.

CaiLama combines your own games, platform profiles, completed online and over-the-board games, engine-checked analysis, key positions, training cards and protected web functions in a controlled test environment.

Analysis quality

Game to training

  • Own PGNs plus Lichess and Chess.com profiles provide profile-bound training material.
  • Completed online and over-the-board games enter analysis after the game.
  • Game analysis and result checks run as a controlled workflow.
  • Position and move data feed training cards and feedback.
  • Tasks stay bound to the player profile and training history.
  • Quality checks remain transparent for testers.
Privacy

Profile frame

  • Data export, confirmed deletion of profile data and clear retention rules are part of the profile frame.
  • Platform data, imported games, training cards and reviews stay profile-bound.
  • No private games or comments are used for model training without explicit permission.
Training

Training value

  • Training cards formulate a concrete question for the position.
  • Repetition considers solved and unsolved tasks.
  • Coaches can translate student games into exercises more deliberately.
  • The test phase checks which hints actually help players.
Quality assurance

Systematic measurement

  • Analysis quality, clarity and training value are reviewed regularly.
  • Tester feedback flows into cards, comments and prioritisation.
  • Automatically detectable errors are evaluated separately from human feedback.
  • Public statements stay limited to the current test status.
AI boundaries

Responsible use

  • AI hints stay behind quality, privacy and approval boundaries.
  • Test, evaluation and training data remain strictly separated.
  • Only approved, synthetic or anonymised data is allowed.
  • AI support remains replaceable and bound to the protected test environment.