Skip to main content

Gemini’s Generate Content API Now Stores Requests by Default in New Paid Projects

2 min read

Google changed the logging default for new paid Gemini API projects on October 26. Existing projects retain their settings; developers can opt out.

Gemini’s Generate Content API Now Stores Requests by Default in New Paid Projects

Google changed the Generate Content API’s storage default on October 26, 2026. New paid-tier projects now store requests by default, according to the Gemini API release notes. Existing projects keep their current setting. The change makes logging a configuration decision developers should address when creating a project.

The affected methods are generateContent, streamGenerateContent and batchGenerateContent. Google documents two ways to opt out: send store=false on an individual request, or disable Generate Content API logging in the Settings panel of AI Studio’s Logs and Datasets dashboard.

The project boundary matters

Project or requestDocumented behaviour
New paid-tier projectRequest storage is enabled by default.
Existing projectThe current setting is retained.
Request with store=falseOpts that request out of Generate Content storage.
Scope of Google’s October 26, 2026 change. Check the setting in the project that actually serves the application.

A team migrating an application into a newly created project could therefore encounter a different default while using the same model and similar application code. Copying an old integration example is insufficient evidence of the new project’s logging policy. The relevant configuration belongs in the deployment review alongside the project ID and billing tier.

Private logs and training contributions are separate

Google’s logging documentation describes project-private logs with a default retention window of 55 days. Available windows are 7, 14, 28, and 55 days. Saving logs into a dataset changes that lifecycle: datasets do not expire under the same log-retention window.

Logging does not, by itself, mean that prompts are automatically contributed for model training. The logs and datasets policy distinguishes private project storage from optional dataset sharing and feedback. Abuse-monitoring records are also governed separately. Disabling this logging feature should not be described as erasing every record Google may hold.

For applications handling customer material, the practical question is which requests may be stored, who can access them, and how long each copy remains. That review is separate from selecting a Gemini feature, including connected Sheets and BigQuery data. This report compares Google’s published documents; it does not establish an account-level test or a compliance determination.

Leave a comment

Your email address will not be published. Required fields are marked *