AI usage
Open Translation statistics → Model usage to see AI services and models, request counts, successes, and usage trends.
Open Translation statistics
Check usage
- Choose a time range to view request and usage trends.
- Select a service or model to filter its records. Select it again to clear the filter.
- Filter request records by scenario, status, or cache to inspect failures and retries. You can also export them.
A token is a unit of text processed by a model, not a character or word count. Input tokens come from the text sent; output tokens come from the response. Models can count them differently. When the service supplies details, the page separates input, output, and cache usage.
Read the numbers
Records cover AI requests connected to usage tracking. Some connection checks, retries, and multi-part tasks also count. Missing provider reports, untracked features, and activity on other devices may be absent. Use your provider’s bill for actual charges.
Output speed (token/s) uses successful requests’ output tokens and duration, including waiting and transfer time. Missing required data displays —; zero output displays 0. Average duration still includes all calls.
Save or clear records
Records stay in this browser and can be exported or cleared on the page. They are usage history, not a full translation backup. Clearing them does not reverse provider charges.
Translate only the text you need and limit optional AI context to reduce usage.