Context Window Budget

Made for the task at hand.

Check whether a prompt and its expected response fit a model’s context before making a request. Local o200k_base tokenization supplies a real count, while the budget report separates input, reserved output, and remaining space.

How to use

1. Paste the prompt whose token footprint you want to measure.
2. Enter the context window and reserved output as two comma-separated token counts.
3. Review the fit status, remaining budget, and tokenizer mapping caveat.