AI & Engine Disclaimer — Shogi Chess AI

Effective date: 1 October 2026

1. What is automated here

• Game analysis and evaluations come from chess engines (local Stockfish-family engines, Maia, Berserk, Lc0 where enabled, and the free Lichess cloud service). Scores are estimates produced under time/depth limits and may differ between engines, depths, and runs.
• Explanations, stories, plans, and classroom dialogue are generated by large language models (via OpenRouter) and by text-to-speech services. They can contain mistakes, including plausible-sounding but wrong moves, rules, or advice.

2. Educational purpose only

Everything automated in the app is a training aid, not professional advice — not coaching certification, not academic assessment, and not a guarantee of rating improvement or tournament results.

3. Your responsibility

• Verify anything important (rules, opening theory, critical game decisions) against trusted sources or a human coach.
• Do not present AI-generated content as your own certified work where that would breach competition, school, or federation rules.
• Move classifications (“brilliant”, “blunder”, …) are heuristic labels from score thresholds, not objective judgments of your play.

4. Availability and limits

Free AI allowances, cloud evaluations, and voice features are limited and depend on third-party providers; they may be slow, degraded, or unavailable. Analysis quality scales with the depth/time you select.

5. Feedback

If the AI says something wrong or unsafe, tell us at ravi.sleepnow@gmail.com or utthamchess@gmail.com so we can improve prompts, filters, and fallbacks.