Book Free Demo WhatsApp Us

Token

Book a Free Demo

We'll only use these details to contact you about the training. No spam.

Quick Answer

A chunk of text, roughly a word or part of a word, that an LLM processes as a unit - the basis for both its context window and usage-based pricing.

Language models don’t process text in whole words - they break it into tokens, chunks of text usually smaller than a full word, reading or generating one at a time. Every prompt you send and response you get is measured in tokens, and most AI providers price their models per token.

Example

The sentence “n8n automates workflows” might break into four or five tokens depending on the model’s tokenizer. A model with a 128,000-token context window can hold roughly that many tokens of conversation, documents and response combined before older content starts dropping out.

n8n relevance

Building AI workflows in n8n, token usage affects both cost and what fits into a single request - a long document dropped straight into a prompt can blow past a model’s context window, which is part of why RAG exists: it passes only the relevant excerpt rather than the whole document. Keeping prompts tight also keeps API costs predictable on a workflow that runs often.

Related concepts

See also prompt, RAG, and AI agent.

← Back to Glossary

Call WhatsApp Free Demo