Repository navigation
VRCPulse v2.0.1
Fixes
- Fix AI InsightCard incorrectly showing in the desktop app (should be web-only)
- Fix Gemini API endpoint (v1 → v1beta) for structured output support
- Fix current metric value using 5-minute average instead of single data point for more stable analysis
- Fix rate limit handling with 120s automatic backoff
Changes
- Upgrade AI model from Gemini 2.0 Flash to Gemini 2.5 Flash
- Change cold start data check from 1-hour to 24-hour window for more accurate baseline
- Inject
GOOGLE_AI_STUDIO_API_KEYvia GitHub Secrets in deploy workflow
Full Changelog: v2.0.0...v2.0.1