What is a token in AI?
A token is a small chunk of text, often a whole word or part of one, that language models read and write and that providers use to measure usage.
Language models do not read text letter by letter or strictly word by word. They split it into tokens, which are common chunks of characters. A short, common word is often a single token, while a long or unusual word may be split into several pieces, and punctuation and spaces count too. Numbers, code, and some languages can take more tokens to express the same amount of meaning.
Tokens matter to everyday users in two ways. First, limits: the amount a model can handle at once, its context window, is counted in tokens, so a very long document may not fit. Second, cost: when AI models are used through an API, such as inside an automation, charges are usually based on the number of tokens sent in and generated. An automation that summarizes every incoming email therefore costs roughly in proportion to how much text passes through it.
There is no fixed conversion between tokens and words, because it depends on the model and the text. Tools that show token counts or estimates are the reliable way to check. For planning, it is enough to know that longer prompts, longer documents, and longer replies all use more tokens, and that different models split the same text differently, so their counts are not directly comparable.
An example
An automation that summarizes long customer emails uses many more tokens than one that only sorts short messages by topic, which matters when usage is billed by the token.