Small Utility Lab

CHECK TEXT SIZE LOCALLY

Estimate text size before sending it to an LLM.

Count words and characters, then calculate a clearly labeled rough token estimate from an adjustable assumption. Use the official tokenizer for a model-specific count.

Up to 100,000 characters. Text is processed in this browser tab.

This adjustable ratio is an assumption, not a tokenizer rule.

Enter text to estimate its size.

Estimated counts will appear here.

Method and limits

The word count splits text at whitespace. The displayed token estimate multiplies that count by the adjustable tokens-per-word assumption and rounds to the nearest whole number. The default 1.3 is only a starting assumption; it is not a measurement or a claim about any model.

Tokenization depends on the tokenizer and text. For a more precise count, choose the relevant model/tokenizer in the OpenAI tokenizer. This estimator does not identify a model, count special tokens, or reproduce that tokenizer.

Text stays in this browser tab and is not uploaded or saved by this tool. Inputs over 100,000 characters are rejected. Page content updated November 3, 2026; source link checked the same day.

Source

The official tokenizer is provided for model-specific token inspection. Page content and link checked November 3, 2026.

How it works

1. Paste or type text. 2. Review the whitespace-separated word and character counts. 3. Adjust the tokens-per-word assumption to see a rough estimate; use a model-specific tokenizer for a precise count.

Good to know

Is this an exact token count?

No. It multiplies a whitespace-separated word count by your adjustable assumption. It does not run a model tokenizer.

How is the estimate calculated?

The estimate is the word count multiplied by the assumed tokens per word, rounded to the nearest whole number. The default is an editable assumption, not a measured conversion rate.

How can I get a more precise count?

Use the official tokenizer and select the relevant model or tokenizer when available. Tokenization varies with tokenizer and text.

Does my text leave this browser?

No. The estimate runs locally in this browser tab. Text is not uploaded or saved by the tool.

Is it free?

Yes, during beta. Optional usage-credit plans are coming soon. Payments are not available and no credits are required.