Reports

Reports holds a set of pre-built dashboards over the same trace and observation data you can browse in Observability, aggregated so you see patterns rather than individual requests.

Three ship with the platform.

• Cost breaks spending down by model, environment, user, and use case, with percentile costs per trace and per observation to separate typical spend from outliers.

• Usage Management tracks volume: traces, observations, and numeric and categorical scores, over time and by environment.

• Latency covers response time by use case, level, model, and user, alongside time to first token and output throughput per model.

A date range and filters apply across a whole dashboard, and individual widgets can be downloaded on their own.

Reports aggregates LLM activity across your project. For a live snapshot scoped to selected agents, see Home. To inspect one request, see Observability. For container traffic and resource use, see Analytics.