Costs & Limits
1 min read
Monitor and control AI model expenses with real-time tracking, spending limits, and automatic optimizations.
Statistics
Track usage and costs across teams, profiles, and models with time-based filtering (hour to 12 months).
Key metrics:
- Team costs with member/profile counts
- Individual profile usage
- Model breakdown by cost percentage
- Interactive charts for trend analysis
Usage Limits
Set spending limits to prevent budget overruns:
LLM Limits
- Apply to organization, teams, or profiles
- Daily/monthly reset periods
- Actions when limit reached (block, alert, fallback)
Auto-cleanup
- Configure data retention (hourly to monthly)
- Keep costs database optimized
Optimization Rules
Automatically switch to cheaper models based on conditions:
Rule Types:
- Content Length - Use cheaper models for short prompts (<500 tokens)
- Tool Presence - Simpler models when no tools required
- Time-based - Off-peak optimizations
Rules apply by priority order with configurable target models.
Integration
OpenTelemetry Export
Export metrics to Prometheus, Datadog, New Relic
REST API
GET /api/costs/statistics?period=7d
GET /api/costs/limits
POST /api/costs/optimization-rules