Overview#
RadarB3 is a full-stack quantitative analysis platform for the Brazilian and global financial markets. It brings together quotes for B3 assets, cryptocurrencies and US stocks, email alerts, backtesting and AI-generated insights in a single terminal. It is in development.
"Where strategy meets technology."
Problem#
Quantitative traders usually juggle several tools: one for B3 quotes, another for crypto, another for charts and indicators, and spreadsheets to test strategies. Data is scattered and alerts depend on watching the screen.
Solution#
- Dashboard with a consolidated market view, watchlist and key metrics
- Screener to filter assets by technical and fundamental criteria
- Charts with candles, RSI(14) and moving averages across several timeframes (15m to 1W)
- Price alerts with history and email notifications
- Backtesting of strategies on historical data
- AI insights with Google Gemini
Architecture#
- Next.js App Router holds both pages and API routes (
auth,market,alerts,backtest,health) - Cascading data sources: BRAPI for B3, Binance with CoinCap fallback for crypto, AlphaVantage for US stocks
- TTL cache in Postgres for market data
- Graceful fallback: without the optional keys, the platform keeps working with mock data
- Rate limiting per IP and structured logging
Tech Stack#
- Web: Next.js 16, React 19, TypeScript, Tailwind CSS 4, Motion, Recharts
- Data: Neon (PostgreSQL) via
pg, Supabase as an optional auxiliary provider - Auth: JWT and bcrypt
- Integrations: BRAPI, Binance, CoinCap, AlphaVantage, Google Gemini, Resend
- Deploy: Vercel
Status#
In development. Next steps listed in the repository include automated tests, payload validation with Zod, WebSocket quotes and versioned migrations.

