Screener Data Architecture - FMP/Finnhub/Yahoo Cascade

Data Sources (Priority Order)

1. FMP (Financial Modeling Prep) - Primary

   • New API: /stable/quote, /stable/profile, /stable/ratios-ttm
   • Free tier: 250 calls/day
   • Best for fundamentals

2. Finnhub - Fallback

   • Free tier: 60 calls/min
   • Good for real-time quotes

3. Yahoo Finance - Last resort

   • Via yfinance library
   • No API key needed
   • Rate limits apply

Stock Universe

~200 stocks: S&P 500 + Nasdaq 100 (deduplicated)
~50 ETFs: SPY, QQQ, sector ETFs, leveraged

Architecture

• Daily cron pulls data → Supabase screener_stocks table
• Frontend reads from cached data (fast, no API calls)
• Cascade fallback when rate limited

Endpoints

• GET /api/screener - Get cached screener data
• POST /api/screener/sync - Trigger data sync
• GET /api/screener/status - Check data freshness
• GET /api/screener/symbols - List tracked symbols

FMP API Notes

• Old v3 endpoints deprecated Aug 2025
• New /stable/ prefix required
• Batch quotes NOT available on free tier