AI Disclaimer
Important information about the use of AI-powered features
1. Nature of AI Outputs
Chess Trainer uses the large language model (LLM) Claude by Anthropic to generate chess analyses, coaching hints, and player profiles. These outputs are automatically generated suggestions and do not constitute professional chess advice.
AI models can produce factually incorrect, incomplete, or misleading content – even for seemingly clear positions or recommendations. All AI outputs should be evaluated critically and not used as the sole basis for decisions in tournament or competitive play.
2. AI Technology Used
Claude (Anthropic)
The AI coaching features (chat analysis, player profile) are powered by Claude Sonnet 4.5 by Anthropic PBC. Claude is an LLM trained using deep learning that generates text that appears plausible and helpful but is not necessarily factually correct.
Stockfish
Move-by-move evaluation (blunders, mistakes, inaccuracies) is performed by Stockfish – an open-source chess engine that operates deterministically and according to established chess principles. Stockfish outputs are significantly more reliable than LLM-generated text but may vary depending on search depth and position complexity.
- stockfishchess.org (GNU GPL v3)
3. Third-Party Libraries Used
Chess Trainer is built on a number of open-source libraries and third-party services:
Frontend
- Next.js \u2013 React framework (MIT License)
- React \u2013 UI library (MIT License)
- Tailwind CSS \u2013 CSS framework (MIT License)
- shadcn/ui \u2013 UI component library based on Radix UI (MIT License)
- Radix UI \u2013 Accessible UI primitives (MIT License)
- react-chessboard \u2013 Chessboard component (BSD-3-Clause)
- chess.js \u2013 Chess logic (BSD-3-Clause)
- Recharts \u2013 Chart library (MIT License)
- react-markdown \u2013 Markdown renderer (MIT License)
- Lucide React \u2013 Icon library (ISC License)
- TanStack Query \u2013 Data synchronization (MIT License)
Backend
- FastAPI \u2013 Python web framework (MIT License)
- python-chess \u2013 Chess library for Python (GPL v3)
- Pydantic \u2013 Data validation (MIT License)
- SQLAlchemy \u2013 ORM (MIT License)
- Anthropic Python SDK \u2013 Claude API client (MIT License)
Infrastructure
- Supabase \u2013 Database & authentication (Apache 2.0)
- Stockfish \u2013 Chess engine (GNU GPL v3)
- Lichess API \u2013 Chess game import (AGPL v3 / public API)
- Chess.com API \u2013 Chess game import (public API)
4. No Warranty for AI Quality
AI-generated analyses and recommendations are provided without any warranty. In particular, there is no guarantee that:
- chess move evaluations by the AI coach are factually correct,
- the generated player profile captures all game-relevant patterns,
- recommendations are appropriate in tournament or competitive situations.
We recommend treating AI outputs as supplementary learning inspiration rather than authoritative chess expertise.
5. Data Processing by Anthropic
When you use the AI coach or profile generation, parts of your chess data (move sequences, positions, analysis results) are transmitted to the Anthropic API. By using these features, you consent to this transmission. Anthropic processes this data according to its own privacy policies.
You can choose not to use these features at any time without affecting core functionality (game import, Stockfish analysis).
Last updated: 19/02/2026