Analytics

The analytics dashboard gives you visibility into how your team uses AI coding assistants. Track costs, token usage, and agent performance over time.

Summary KPIs

The summary view shows headline metrics for a selected time period:

• Total sessions — Number of coding sessions
• Total cost — Estimated spend across all sessions
• Total tokens — Input + output tokens consumed
• Average session duration — Mean session length
• Average cost per session — Mean cost
• Unique users — Number of active team members

Timeseries

The timeseries chart shows metrics over time. Toggle between:

• Sessions per day/week
• Cost per day/week
• Tokens per day/week

Use the date range picker to focus on specific periods.

By agent

Break down metrics by agent to see which agents get the most use and which perform best:

• Session count per agent
• Average cost per agent
• Average duration per agent
• Token usage per agent

This helps you identify which agents to invest in improving.

Cost breakdown

See where your spend goes:

• By model — Cost split across Claude models (Opus, Sonnet, Haiku)
• By project — Cost per project
• By user — Cost per team member

Per-user analytics

Managers can view per-user metrics to understand individual usage patterns:

• Sessions per user over time
• Cost per user
• Agent preferences
• Session duration patterns

Filtering

All analytics views support filtering by:

• Date range — Custom start/end dates
• Project — Filter to a specific project
• Repository — Filter to a specific repo (for workspace projects)
• Agent — Filter by agent type
• User — Filter by team member

API access

All analytics data is available via the API for building custom dashboards or integrations:

GET /:org_id/sessions/analytics/summary
GET /:org_id/sessions/analytics/timeseries
GET /:org_id/sessions/analytics/by-agent
GET /:org_id/sessions/analytics/cost-breakdown

See the Sessions API reference for details.

Next steps

• Team management
• Experiments