Cost per prompt
API-equivalent USD for your 12 costliest prompts · Lower is better
Your Claude Code, analysed
Every prompt, token and API-equivalent dollar on this machine
LIVE
Connecting…
Auto-refreshing every 10 s
Speed
Output tokens per second, median per request · Higher is better
Tokens per prompt
Total tokens incl. cache for your 12 heaviest prompts · Lower is better
Cost
What your usage would cost at Anthropic API list prices, split by token type. Cached reads are cheap; writes and output are not.
Tokens & Cache
Input, output, cache writes and cache reads, over time and per prompt.
Speed & Response time
Response time is the time from the previous event to the end of the model's reply; it excludes tool run time.
Models
Which model, effort level, project and Claude Code version did the work.
Prompts
Everything you typed, with what it cost. Click a row to read the prompt and its requests.
All prompts
Tools
Which tools Claude called, counted per request.
Sessions
One row per Claude Code conversation. Click a row to filter the page to that session.
All sessions
Requests
Every API call, token by token. Hide or show columns with the Columns button.