back to projects
Maestro
20264 technologies

Maestro

ai-powered conversational football assistant

year

2026

stack
NextJsTypeScriptGemini APISupabase

about this project

an llm-orchestrated football assistant with 19 function declarations. users ask questions in natural language; gemini 2.5 flash decides which live-data apis to call across football-data.org, api-football, thesportsdb, and newsapi, executes them concurrently via promise.all, and returns combined text + visualization responses through 19 typed ui components. features in-memory ttl caching (30s-3600s), fuzzy team name resolution with ufd normalization, 3-tier personalized fixture recommendations, and multi-turn conversations stored in supabase.

built with
NextJs
TypeScript
Gemini API
Supabase