Glossary · AI
Token
Also called: AI token, LLM token, input and output tokens
A token is the unit of text a language model reads and writes, usually a word, part of a word or a punctuation mark, and the unit AI providers use to price usage and set context limits.
Models don't see characters or words; they see tokens from a fixed vocabulary. Common words are often one token, rare words split into several, and numbers, code and non-English text usually take more tokens than you'd guess. Each provider's tokenizer differs, so the same text has a slightly different count on each.
Why tokens matter for automation
Cost. Providers price input tokens (your prompt, instructions and data) and output tokens (the model's reply) separately, and output is typically priced higher. A per-run estimate is simply:
cost ≈ (input tokens × input price + output tokens × output price) ÷ 1,000,000with prices quoted per million tokens on the provider's pricing page.
Limits. Every model has a context window: the maximum tokens it can consider at once, input and output together.
Practical ways to spend fewer
- Send only the fields the step needs, not the whole record.
- Cap output length, and ask for structured output instead of prose when software reads the result.
- Use a smaller model for classification and extraction; save large models for hard reasoning.
- Don't put a long, unchanging policy document in every call if a short summary does the job.
In Fluxo Flow
Every AI step in a run trace shows the tokens it used and an estimated cost, from list prices for popular models (your provider's bill is the source of truth). You set a max-tokens limit per step, and each server run stops if its estimated AI cost passes a per-run cap. Dry Runs never call your provider, so testing uses no tokens at all.
Read next
Related terms
- LLMAn LLM, or large language model, is a neural network trained on large amounts of text that generates text by repeatedly predicting the next token, which lets it summarize, classify, extract and draft.
- Context windowA context window is the maximum number of tokens a language model can take into account in a single request, counting the instructions, the input data and the reply it writes.
- BYOK (bring your own key)BYOK, or bring your own key, means a software product runs its AI features on an API key from your own AI provider account, so usage is billed to you directly by the provider instead of being resold through the product.
- Structured outputStructured output is a model response in a predefined, machine-readable format, typically a JSON object matching a schema, so the next steps of a workflow can use the fields directly instead of parsing prose.
- TriggerA trigger is the event that starts a workflow run: a person pressing Run, a schedule coming due, a webhook arriving, or a new record or event in a connected app.
- Sub-workflowA sub-workflow is a workflow that another workflow calls as one of its steps, passing data in and getting a result back, so shared logic is built once and reused.