Usage
Track AI token consumption for Beaver Builder AI generation activity, broken down by time period and model.
Opening the Usage tab​
The Usage tab tracks AI token consumption for all generation activity on your site. It is a tab on the Beaver Builder AI settings page, visible to WordPress administrators only. In the WordPress admin, go to Settings > Beaver Builder AI and select the Usage tab.
Reading the usage dashboard​
At the top of the page, choose a time range: 7d, 30d (default), or 90d. The selected period's start and end dates appear with the results.
Summary metrics​
Two totals are displayed for the selected period:
- Input Tokens. Total tokens sent to the AI provider in prompts and context.
- Output Tokens. Total tokens generated by the AI in responses.
The dashboard does not display cost figures. Pricing varies by model and changes over time, so for accurate costs, check the usage console of your AI provider (Anthropic, OpenAI, or OpenRouter).
Per-model breakdown​
The breakdown table groups activity by model, with one row per model used in the period:
| Column | Description |
|---|---|
| Model | The model identifier, for example claude-sonnet-5 or gpt-5.6-terra. |
| Requests | Number of API calls made to this model. |
| Input Tokens | Input tokens consumed by this model. |
| Output Tokens | Output tokens generated by this model. |
If you switch providers or models during the period, each model appears as its own row.
What counts toward usage​
Usage on this page reflects prompts sent through the built-in AI chat in Beaver Builder and the block editor, which run on the API keys entered in Settings.
MCP-connected agents such as Claude Code or Cursor run on their own model and billing. Their activity does not consume your site's API keys and does not appear on this page.
Keeping costs predictable​
A few habits keep AI spending under control:
- Set a monthly spending limit in your AI provider's billing dashboard.
- Use the 30-day and 90-day views to identify periods of high usage and correlate them with specific generation activity.
- Break large requests into smaller, focused prompts. Generating one section at a time uses fewer tokens per request than regenerating a full page.