Repository navigation
Usage and Performance
Paulo Maia Carvalho edited this page Oct 2, 2026
·
1 revision
The Usage tab shows what the agent is consuming while you work: tokens, cost, how well prompt caching is working, and an estimate of the energy that caching saves.
- Totals: total tokens, session cost, number of agent turns and the cache hit rate
- Token breakdown: a stacked bar for the whole session and for the last turn, split into input, output, cache write and cache read tokens
- By provider / model: a table that groups turns by provider and model, with summed tokens and cost and the average cache hit rate
- Turn history: the most recent turns (up to 50), kept in your browser so they survive a page refresh. A clear button empties the history
- Environmental impact: an estimate of the CO₂ and water saved by prompt caching, and roughly how many tree-days that equals
Prompt caching lets the provider reuse the part of the conversation it has already processed instead of processing it again. A high hit rate means lower cost and less compute for the same work. The environmental figures are estimates derived from the tokens served from cache, not measurements.
- The numbers come from the agent context reported by PiClaw, so they reflect the provider and model you are using
- Turn history is stored per browser. Opening DrupalClaw from another browser starts with an empty history