---
title: "Viewing AI Usage"
canonical: https://trickest.com/docs/using-the-app/agent/viewing-ai-usage
description: "Open Billing’s Usage tab and distinguish AI activity, budget and request limits."
---
# Viewing AI Usage
## Open AI usage
Open **Settings > Administration > Billing**, then select **Usage**. The page title is **Billing**; AI activity appears inside its Usage tab.
You need AI access to see the AI sections. A user with plan-management access may also see **Plan limits** and the **Plan** tab. If you have AI access without plan access, Billing opens Usage; the Plan tab can be absent.
| Tab | What to inspect |
| --- | --- |
| **Plan** | The vault’s plan and subscription controls, when you have permission to manage them. |
| **Usage** | Plan limits for permitted plan managers, and AI usage for users with AI access. |
For model access, see [Managed AI](/docs/key-concepts/agent/managed-ai) or [Bringing Your Own Key](/docs/using-the-app/agent/bring-your-own-key).
## Read the budget
When the vault has a current AI grant, **AI budget** shows the amount spent from that grant and the remaining budget for its displayed period. **Limit reached** means the applicable budget or cap has been reached. **Near limit** indicates that it is close.
**No active contract period** means the configured grant does not cover the current time. Read the accompanying start or end date instead of treating it as a fresh zero-spend budget.
A section titled **Your usage limit** applies to your own usage. It is separate from a vault-wide AI grant. Budget and limit sections appear only when applicable data is available.
## Read activity
**AI usage** shows **Tokens** and **Requests** for the displayed period. Tokens measure model activity; they are not the remaining dollar budget.
**Trickest spend** is spend billed through Trickest. Calls made with your own provider keys are billed by that provider, and Trickest spend excludes those charges. The page can identify **Provider keys**, **Platform** or **Mixed billing** usage. Check your provider’s dashboard for provider-key charges.
**By model** breaks down the vault’s activity by model. Its token share describes usage distribution, not the model’s quality or the budget remaining.
## Limits
Some vaults show **Vault limits** or **Your limits**, depending on the applicable configuration. Check the section’s scope and notes. **Disabled** limits do not block requests; other enabled limits or the vault AI budget can still apply.
When an enabled blocking limit is reached, new AI requests stop for the scope stated on the page until the limit resets. A vault-wide limit and a personal limit can have different effects. Vaults using verified provider keys may omit the limits section.
## If usage is unavailable
If Billing or Usage is missing, confirm the selected vault and your AI or plan-management access. A missing Plan tab alone does not mean you lack AI access.
Use **Retry usage** or **Retry limits** when the page offers it. A refresh warning can accompany previously loaded values; read that warning before treating those figures as current. An unavailable or empty usage section is not proof that no requests were made.
---
_Markdown source of https://trickest.com/docs/using-the-app/agent/viewing-ai-usage._