Usage and Analytics

Each project has a dashboard area showing what's happening across your integration: how much it's being used, how much it's costing, and how well the agent is doing at completing what users ask it to do.

What you'll find here

• Volume — message and session activity over time, so you can see how usage trends as your app grows or as you ship changes.
• Spend — token and usage consumption, viewed against any limit you've configured. See Permissions, budget, and personalization for setting that limit.
• Agent performance — a higher-level view of how well the agent is completing user requests, including successful versus failed or abandoned interactions. This is the place to spot problem screens or actions that might be worth improving — see MCP metadata quality and Agent actions for how better metadata tends to translate into better completion rates.

Exact metric names and chart layouts aren't part of the public contract and may evolve — treat this page as a guide to what kind of information is available, not a fixed reference. There's no separate public metrics API; this data is only available through the dashboard.

Next steps

• Permissions, budget, and personalization — where spend limits are configured.
• MCP metadata quality — improving how well the agent understands your screens and actions.
• Concepts: Agent actions — what the agent's actions look like under the hood.