Skip to main content
The Analytics section (/analytics) gives you full cost visibility across the platform. All data is sourced from the Dashboard API analytics module and updates reactively as new token events arrive.

​
Cost Explorer

Interactive cost analysis with filtering by gateway, agent, model, workflow, or time range. The explorer shows token spend over time as a filled area chart with configurable bucketing (hourly, daily, weekly).
ActionHow
Filter by dimensionSelect Agent, Model, Gateway, or Workflow grouping
Drill downClick any bar or segment to filter the view to that dimension value
ExportDownload the current view as CSV for external analysis
Time rangeSegmented control: 24h, 7d, 30d
Click through to individual token events for per-request detail — see the exact model, token counts, cost, and associated workflow run.

​
Model Economics

Per-model cost comparison to help identify the most cost-effective models for each use case.
MetricDescription
Input vs output pricingPer-million-token rates for each model
Average cost per requestMean cost across all requests using this model
Volume distributionHow token volume is distributed across models
Substitution opportunitiesPotential savings from switching models, with estimated dollar impact
Substitution suggestions are directional estimates based on pricing differences. Quality and capability trade-offs are noted but not quantified.

​
Budget Tracking

Current spend vs allocated budget at each hierarchy level: company, department, team, worker.
Under budget

Green progress bar, under 80% consumed

Approaching limit

Amber progress bar, 80-100% consumed

Over budget

Red progress bar, budget exceeded

Each budget level shows a progress bar with percent consumed, projected overage warnings, and a top-cost-drivers list. Budget thresholds can be set for daily, weekly, and monthly periods with enable/disable toggles.

​
Alert Rules

Configure cost threshold alerts per gateway from the Settings page, or use the built-in automatic alerts:
ThresholdLevelCondition
Daily fleet cost exceeds $5Warning24h total spend
Daily fleet cost exceeds $20Critical24h total spend
Daily per-agent cost exceeds $2WarningAny single agent’s 24h spend
Alerts appear in the notification bell and (if configured) in Slack. Deduplicated within each hour to prevent alert storms.

​
Token Event Detail

Drill into any token event to see the full picture:
FieldDescription
ModelWhich model processed the request
AgentThe agent that generated the tokens
GatewayWhich gateway handled the request
TokensInput, output, and cached token counts
CostDollar cost for this event
StatusSuccess or failure
RunLink to the associated workflow run
How it works: See Token Spend Attribution for the full attribution chain. →