>_ PromptCostt

Paste a prompt (and optionally attach a text file it depends on), and PromptCostt counts the input tokens right in your browser and multiplies them by each model's published price. Set how many times a day you expect to run it to project daily and monthly spend.

Because a response's length is unknown in advance, each model shows a min, expected, and max cost — a short reply, a typical reply, and your own response cap. For the most predictable costs, state the cap in the prompt itself (e.g. "Respond in at most 500 tokens"). These are rough estimates for comparing models and ballparking a budget, not an exact invoice — see the notes at the bottom for details.