Playground

A tiny writing lab for prompts

This page is intentionally small. It is not a full-featured tokenizer. It is just enough UI to notice that when your writing gets more polite, more structured, or more verbose, the token count quietly goes up.

146 chars b7 ~37 tokens

What the model feels

Rough estimate only; real APIs use provider-specific tokenizers.

Length

~37 tokens

Every polite sentence, bullet list, or extra example lives here.

Shape

6 lines

Models do not see paragraphs, but structure still nudges behavior.

Tradeoff

cost 22 tokens

Try rewriting the same ask in fewer words and notice how the estimate shifts.

If you want real numbers, use your provider's tokenizer API. This page is here to build a small habit: when you tweak a prompt, glance at the length and ask yourself whether the extra words are buying you better behavior or just a higher bill.