Token
The unit of text a language model processes, typically a word or word fragment, used to measure both context limits and usage-based pricing.
In more detail
Models do not read characters or words directly; they read tokens, which are chunks of text roughly the size of a short word or word fragment. A sentence becomes a sequence of tokens on the way in, and the model emits tokens on the way out. Token counts, not character counts, are what matter technically.
Two practical consequences follow. Context windows are measured in tokens, so the token count of the input decides whether it fits. And most model pricing is per token for input and output, so the token footprint of a workflow is a direct driver of its running cost.
Where this shows up at Ceven
Because tokens drive both limits and cost, Ceven keeps workflow steps tight: passing a step the material it needs rather than everything available. That discipline is part of how a workflow you describe in plain language stays affordable to run repeatedly, and why usage stays legible in the audit trail.