- Add POST /api/quotes/openai/stats/update endpoint that pings OpenAI API
- Update fetchActualUsageFromAPI() to make real API call for fresh data
- Add gpt-4o-mini pricing support in calculateTokenCost()
- Improve frontend update button with loading state and real-time data
- Replace static token counts with live tracking from OpenAI ping requests
- Reduce cache time to 1 minute for more responsive updates
- Add proper loading states and error handling for budget updates
alex
8c181be737