# Usage

Usage shows the tokens that the AI features of your workspace consumed, recorded per model. It is under [**Usage**](https://auth.faustine.ai/go?to=%2Fusage) in the portal's sidebar. Faustine does not charge for this usage. Your model provider bills it to you directly.

## How it works

Faustine records the tokens each model consumes during analysis, chat, the Advanced editor, voice and automatic updates. The page adds them up for the range you select: **7d**, **14d** or **30d**.

The amounts in dollars are estimates and are written with a tilde, such as `~$12.40`. They can differ from your provider's invoice. A model without a rate shows its tokens and **no rate** in place of an amount.

## Totals

| Card | What it counts |
| --- | --- |
| **Tokens** | Input, output and cache tokens, added together. |
| **AI requests** | Calls to a model. |
| **Estimated cost** | The estimate for the range. When some requests ran on a model without a rate, a line under the amount says how many are not included. |

## Estimated cost over time

A bar per day with that day's estimate. Requests on models without a rate are not in the bars.

## By feature

**By feature** splits the same usage by the part of Faustine that consumed it. Select a feature to see its sources.

| Feature | Sources |
| --- | --- |
| **Web Agents** | Conversations, Analysis, Auto-updates, AI editor, Test generation |
| **APIs** | Analysis, Updates, Recipes, Recipe imports, Test generation, AI editor |
| **Faustine Chat** | Conversations |
| **Other** | Usage that belongs to none of the features above. |

Each row shows its requests, its tokens and its estimate. A row where nothing could be priced shows tokens only.

## By model

**By model** lists every model with its requests and tokens, and the estimate where there is a rate. Model IDs are shown without their date suffix.

## Troubleshooting

| Symptom | Cause | Fix |
| --- | --- | --- |
| **No AI usage in the last 30 days** | Nothing ran in the selected range. | Select a longer range, or run an analysis or a conversation. |
| A model shows **no rate** | Faustine has no public rate for that model. | None needed. The tokens are recorded, and your provider's invoice has the amount. |
| The estimate differs from the provider's invoice | Estimates use list prices, and your contract with the provider has its own. | Use the provider's invoice for accounting and this page for the split by feature. |

## Next steps

- [Billing](https://docs.faustine.ai/workspace/billing/): what Faustine charges the workspace.
- [Monitoring](https://docs.faustine.ai/workspace/monitoring/): the runs and conversations behind these numbers.
- [Read Faustine Chat logs and usage](https://docs.faustine.ai/apps/chat-logs-and-usage/): usage per chat user.
- [Model providers](https://docs.faustine.ai/workspace/model-providers/): the account these tokens are billed to.