Usage Dashboard
The Usage Dashboard gives you a breakdown of how your Claude Code sessions consume tokens, which tools they call, and what it costs. Open it from the sidebar or via the command palette.
The dashboard is organized into four tabs:
Tokens
Section titled “Tokens”Shows aggregate token counts across all sessions:
- Input tokens — Tokens sent to the model
- Output tokens — Tokens generated by the model
- Cache write tokens — Tokens written to prompt cache
- Cache read tokens — Tokens served from prompt cache
A per-project breakdown table shows how token usage is distributed across your projects.
Displays tool usage metrics with a sortable table and stacked bar visualization:
- Count — How many times each tool was called
- Duration — Total execution time per tool
- Tokens — Tokens consumed per tool
- Cost — Estimated cost per tool
Use the sort selector to reorder by any of these dimensions. Bar charts are colored by model to show which models drive which tool usage.
Skills
Section titled “Skills”Same layout as the Tools tab but for Claude Code skill invocations. Shows which skills are used most frequently and how much they cost.
Cost Estimator
Section titled “Cost Estimator”Estimates your total spend based on token usage and model pricing:
- Time range — Filter by 1 day, 7 days, or all-time.
- Per-model pricing — Editable price-per-million-tokens for input and output. Defaults are fetched from live pricing data and can be overridden.
- Grand total — Calculated cost across all models and time ranges.
The cost estimator is an approximation based on token counts and published pricing. Actual billing may differ depending on your API plan.