TokenPad
Counting

Token Heatmap

See which lines of a prompt are eating the budget.

Prompt or document, one unit per line

209 characters4 lines0 tokensor drop a file

Token HeatmapExact
0Token change
Token change0no change
Input tokens0what you pasted
Output tokens0what you would send
Token cost of this result
Output tokens0
As input$0.00
× 100K requests$0.00

Everything on this page runs in your browser. Nothing you paste is transmitted, because there is no server here to transmit it to.

Result

Density tells you what you are looking at

Characters per token below three means the line is JSON, code, identifiers or non-Latin script. Those lines cost two to three times more per character than the prose around them.

The flag is the fastest way to find the one embedded object or base64 blob that is quietly responsible for half of a prompt.

What to do with a hot line

Minify it if it is JSON, convert it if it is a table, strip comments if it is code, or ask whether the model needs it at all. The prompt optimizer and data format optimizer handle the first two mechanically.

The counts here are estimated from character length so many lines profile instantly. Confirm anything you act on in the token counter.

Frequently asked questions

Why estimate rather than count exactly?
Speed across many lines. The estimate is good enough to rank sections, which is what this tool is for. Use the token counter for the exact figure on whatever you decide to change.
My whole prompt is flagged dense. Now what?
You are probably sending structured data rather than prose. Format choice will beat line-level editing — compare serialisations in the data format optimizer.

More counting tools