Skip to main content

Managing AI Requests and Tokens

Archie tracks the Total AI Requests and Total Tokens Consumed at the bottom of the screen.

A token is a unit of data that is processed by algorithms in natural language processing (NLP) and machine learning services. A token is a component of a larger data set, which may represent words, characters, or phrases. When processing text, a sentence is divided into tokens, where each word or punctuation mark is considered a separate token. The process of tokenization is a step in preparing data for further processing in AI models.

The numbers displayed relate to the project that is currently open.

Total AI requests and total tokens consumed in an Archie project

  • Total AI Requests: The number of API calls to different language models that have been run in your project.
  • Total Tokens Consumed: The total number of tokens this project has consumed. When you run a Propagate or Regeneration, you consume tokens. If your project goes over the token allowance, the credit card used to sign up will be charged. You can click Info to see how many tokens you have left.

Users with Viewer permissions cannot edit projects, and therefore cannot trigger AI requests or consume tokens. If you find that a project is generating too many requests, you can convert some Editors to Viewers. For more information, see Roles.