AI Usage
Usage shows how much work passed through each AI account and model and what made up the latest request.
What the overview shows
Open Usage in the sidebar. The top summarizes today, this month, and loaded history. Other sections break usage down by day, AI account, and model.
For ChatGPT subscription or Assistela Managed, the page may also show available limits, credit, request count, or plan balance. Provider data is always authoritative for billing.
What tokens are
A token is a small piece of text processed by the AI. The count includes your instruction, required context, descriptions of available tools, and the answer. A token is not the same as a character, a word, or a fixed amount of money.
A long document, extensive history, many enabled tools, or thorough work naturally uses more tokens. Use the particular provider’s plan for prices and limits.
The latest AI request
The latest-request detail helps explain what was included for the model: stable setup, memory, temporary information, calendar, Reminders, tools, and recent Chat. It also shows the context limit, remaining room, and possible compression of older history.
This is an inspection view, not text you normally need to edit. Use it when an answer misses important context or usage rises unexpectedly.
When usage is higher than expected
- Check which account and model created it.
- See whether Personal Assistant or Scheduled Tasks run frequently.
- Increase the automatic-check interval if frequent runs are unnecessary.
- Disable unused tool groups and personal sources.
- Attach large documents only to tasks that need them.
- Change advanced routing and context settings one at a time, then check again.
An overly small context can hide required instructions or part of a document from the AI. Reduce unnecessary sources and automatic-task frequency first.