Metrics Available
- Total requests — by day, week, month with trend graphs
- Cost breakdown — by mode, model, and active skill
- Model distribution — which Theo engines handle your requests (pie chart)
- Latency percentiles — p50, p95, p99 response times
- Cache hit rate — percentage of requests served from semantic cache
- Token consumption — prompt tokens vs. completion tokens
- Error rates — 4xx and 5xx response breakdown
Filtering
Filter usage data by:- Date range — custom start/end dates
- API key — see usage for a specific key
- Mode — filter by
fast,think,code, etc. - Skill — see which skills contribute to usage