USEFUL AI NOTES
Token Budget Estimator
Count text with two bundled encodings, reserve reply space and export a context budget.
Loading your tools…
HOW THIS WORKSPACE FITS YOUR DAY
Bring the task. Make something useful.
01
Divide the context
Put instructions, source material and output requirements into named segments. Add or rearrange segments as needed.
02
Choose your budget
Select the encoding documented by your provider, enter its context capacity, and reserve space for the reply and message overhead.
03
Inspect the allocation
See token counts for every segment and the remaining budget. Export JSON or CSV to keep the assumptions beside the result.
Before you begin
Is this the exact number an API will bill?
The text counts are exact for the selected bundled encoding and separately counted segments. Message formatting, tool definitions, images and provider processing can change actual API usage.
Which encodings are available?
cl100k_base and o200k_base run locally in a browser worker. Choose the encoding your provider documents; the app does not guess a model mapping.
What happens if I exceed the budget?
The workspace shows the overage. Invalid capacity or reserve settings block export until they are corrected. A valid but over-capacity report remains useful for planning reductions.
Where is my project saved?
Projects stay in this browser, separated by your signed-in account. Download a project backup before changing devices or removing browser data. The site does not provide cross-device project sync.