[ Usage ]
Every request, itemised.
Model, tokens, what the model cost, what we added, and how much of our cut went to the burn queue.
Your conversation lives in this browser tab. We do not store prompts or replies — only the metering row: model, token counts, cost and time.