Your text0 chars · 0 words

Token boundaries

Alternating shading = one token each. Exact view available for OpenAI models.

Type something above.

Context-window fit

How much of each common context window your text fills (using the selected model's count).

Cost

Prices change constantly, so nothing is hardcoded — put in the rate from your provider and it computes.

Input $ / 1M tokens

Estimated input cost

$0.000000

On accuracy. OpenAI counts use the real tokenizer (cl100k / o200k), loaded once in your browser — these are exact. Claude and Gemini do not publish a client-side tokenizer, so those numbers are a calibrated estimate (treat as roughly ±15%). We won't show a fake "exact" where none exists. The token-boundary view is only shown for models we can tokenize exactly.

How the token counter works

Paste your text and it counts the tokens in your browser. For OpenAI models (GPT-4o, GPT-4, GPT-3.5) it uses the real tokenizer (o200k and cl100k), so the count is exact. Claude and Gemini don't publish a tokenizer that runs in the browser, so for them it shows a calibrated estimate (about ±15%), labeled as an estimate.

Does my text leave my computer?

No. The counting runs entirely in your browser; nothing is uploaded.

What else does it show?

How much of each context window your text fills, from 8K to 1M tokens, and the input cost at the price per million tokens you enter.