See what AI cost for a card or a board

Read the AI calls made for one feature or a whole board, by kind of work, with an estimated cost.

In short

Every AI call made for a card or a board is recorded against it and sorted by kind of work: Design, Plan, Develop, Test and Ask. A coding agent may also report the tokens it spent. The cost is an estimate from a price list. It is not a bill.

Read a card's usage

Open the Feature card, press "More" on its tabs and choose "Usage".
Read the table under "AI usage": "Calls", "Tokens in", "Tokens out" and "Estimated cost" for each kind of work.
Read the "Total" row.

Result: you see what the feature cost so far. The same totals show on Overview as "AI tokens" and "Estimated cost".

Read a board's usage

Open the board's menu, choose "Reports" and then "AI usage".
Choose the "Period".
Read "By kind of work" and "By card".

Good to know

"1 reported by the agent" counts what the coding agent said it spent. It is recorded and estimated, never charged.
"1 call has no price yet" means no price is known for it yet, so the estimate is a little low.
"No AI was used for this card yet." is normal for a Quick card or one built by people.
For the whole workspace, read /help/ai/ai-usage-report.

If something goes wrong

"The AI usage could not be read right now." - try again in a moment.
A figure looks low - the agent reported nothing, or some calls have no price yet.