Hackathon dev builds AI esports coach that sounds like a real one—after endless prompt tweaks, rate limits, and begging APIs to stop being generic

A developer has created an AI coaching tool, Zenith, that analyzes esports match data and provides personalized insights to players through voice responses. Utilizing GRID's esports API, which offers detailed data on professional League of Legends and VALORANT games, Zenith can track player tendencies, draft patterns, and objective timing. The tool was built for a hackathon and required overcoming technical challenges, including API debugging and prompt engineering. With the help of Junie, a coding assistant, the developer was able to integrate AWS Bedrock, Anthropic, and ElevenLabs' turbo model to enable voice output. The system can generate timestamped review agendas, identify patterns, and predict hypothetical outcomes. While the project is functional, its usefulness in real-world coaching scenarios remains to be tested with actual users. The development of Zenith highlights the potential for AI-powered coaching tools to enhance esports team performance, with companies like Cloud9 potentially benefiting from standardized coaching insights.