Self-hosted docs
Cost Controls
Implement budget guardrails and token usage policies to prevent cost overruns.
Budget thresholds
Set per-route and workspace-level budget limits with alert triggers.
- Example: Route-level budget of $100/day with 80% warning threshold
- Example: Workspace-level budget with departmental allocations
- Example: Real-time budget tracking and alerting
Token guardrails
Enforce token usage policies to control model costs and performance.
- Example: Maximum token limits per request type
- Example: Token usage variance alerts for unexpected spikes
- Example: Token efficiency recommendations based on historical data