RaceCast is a prediction-delivery platform for public-racing fans, handling everything from automated race-data collection to prediction delivery and P&L management. It processes 8 venues nationwide and 300+ races a day unmanned.
Data collection, prediction delivery, instant reach, and P&L visualization — the story of launching a new platform, built in-house from zero in 6 months, that satisfies all four axes essential to public-racing prediction delivery at once.
RaceCast Inc. runs a service that delivers race predictions to public-racing fans. Covering 8 venues nationwide and over 300 races a day, it delivers finalized predictions to users just before each cutoff.
The business model is a prediction-delivery subscription. Users receive predictions via Telegram and manage their betting P&L in the app. Delivery freshness and hit reliability drive continued use.
None of these are technology-selection problems; they are business requirements tied directly to delivery freshness, reach rate, continued use, and reliability. A ground-up build required a design that satisfies them all at once.
Against four design challenges, we built the collection, prediction-delivery, instant-reach, and P&L-visualization domains in-house from zero. An automated operating loop anchored on cutoff times delivers predictions with no human intervention.
Next.js (PWA)
React
Tailwind CSS
NestJS
MySQL
Redis / Bull
AWS S3 / SES
Telegram Bot
⚙ Scheduler / Worker
This platform continuously produces business outcomes — delivery freshness, operational efficiency, continued use, and reliability — rather than technical metrics.