Features
Everything you need to understand your AI features in production.
Feature-level visibility
Understand exactly which product features use AI and how they behave. No more guessing which parts of your app are driving costs.
- Requests per feature
- Cost per feature
- Latency and error rate per feature
- Feature-wise comparison
Agentic task & customer tracking
Track multi-step AI workflows end-to-end. Group all LLM calls by task and attribute costs to specific customers for usage-based billing.
- Group LLM calls by task_id
- Attribute costs per customer_id
- See call sequence for each task
- Total cost and tokens per workflow
Cost & usage analytics
Track AI spend using deterministic calculations based on model pricing. No billing reconciliation required.
- Token usage (input/output)
- Estimated cost per request
- Model-wise breakdown
- Cost trends over time
Reliability & errors
Identify failures before users notice. Get detailed error logs and understand which models and features are failing.
- Error rates by feature and model
- Invalid model and request failures
- Recent error logs with details
- Error trends over time
Scaling health & efficiency score
Know if your AI features are ready to scale. Track unit economics with usage vs. cost correlation, plus a composite efficiency score.
- Usage growth vs. cost change tracking
- Net efficiency calculation
- Weighted efficiency score (reliability, speed, cost)
- Account-level health indicators
More capabilities
Real-time dashboard
See metrics update as events flow in from your application.
Latency tracking
Monitor P50, P95, and P99 latency across features and models.
Environment filtering
Separate production, staging, and development data.
Date range filtering
Analyze trends over custom time periods.
Cost projections
Estimate monthly spend based on current usage patterns.
Error details
See full error messages and stack traces for debugging.