Show HN: GymCoach – Self-hosted workout tracker where you bring your own LLM (github.com)

GymCoach, a self-hosted open-source hypertrophy training tracker, has been launched, allowing users to log workouts and receive personalized coaching powered by their chosen large language model (LLM), such as Anthropic Claude or any OpenRouter model. This platform stands out in the fitness tech space because it emphasizes data privacy; users' training data is stored in a local PostgreSQL database rather than on external servers. The tool features a range of capabilities, including detailed logging of exercises, progress tracking through advanced metrics like estimated one-rep max (1RM), and an AI coach that offers weekly debriefs and program suggestions tailored to the user's actual performance data. This development is significant for the AI/ML community as it illustrates the integration of AI in personalized user applications while emphasizing data ownership and control. The system uses a modular design that prioritizes user privacy and customization, enabling users to choose their AI provider without incurring subscription fees. Technical details highlight its robust architecture, which employs Next.js for both frontend and backend, Prisma ORM for database interactions, and Docker for containerization, making it straightforward to deploy and scale. GymCoach not only reflects a shift toward customized AI interactions in fitness tracking but also reinforces the importance of user agency in managing personal data.