Firebase AI Logic: check the app people already use

Start with the version people actually use

If your app uses Gemini through Firebase AI Logic, ask your builder to check the version people have installed. Firebase AI Logic is Google's service for adding AI features to mobile and web apps. A successful test on a developer's computer is only part of that check.

Starting November 2, Firebase will automatically enforce App Check for AI Logic in projects where it is not already enforced. App Check helps verify that a request comes from a genuine app or device. It sends a short proof with the request, called a token.

Google documented a temporary delay option on October 27. Today is the scheduled start of automatic enforcement. Check your project's current setting and request records to see its actual state.

Read the announcement: Firebase: App Check setup and automatic enforcement for AI Logic.

Check the AI feature across released versions

When enforcement is on, Firebase AI Logic blocks requests without valid App Check proof. That includes requests from released app versions that never had the verification system added. Ask which older versions people still use and whether those versions can send the required proof.

This notice concerns the route through Firebase AI Logic. Your builder can confirm whether your app uses it and identify the affected AI features. App Check verifies the app or device; user sign-in and permission rules still need their own checks.

For practice on a developer's computer, Google provides an App Check debug provider, a testing setup that works while enforcement stays on. Released apps need the appropriate production verification setup before people rely on their AI features.

Make any extra time part of a clear plan

For released apps that still need the work, Firebase allows a temporary delay of automatic enforcement. The project administrator can choose an enforcement date in the Firebase console, under Security, App Check and APIs. The latest available date is January 21, 2027.

That is a setting to review for the individual project, not an automatic extension for every app. Google warns that leaving enforcement off keeps the Gemini access through AI Logic exposed to abuse. Agree a migration plan with the person responsible for the app, including how users will receive the updated version.

After an update, inspect requests from real released versions. Google recommends monitoring how many are verified before deciding that the transition is ready.

Practise the questions you want answered

You can understand this change without implementing it yourself. Ask for a short explanation and a demonstration using approved test information.

1. Confirm whether the AI feature uses Firebase AI Logic.
2. List the released versions people still use and who manages updates.
3. Check the App Check setting and recent verified requests.
4. Agree how older versions will be updated and how the result will be checked.

Bring the same habit into your next build

Our perspective is that a useful app handover should explain how the tool stays dependable as services change. If you are learning, practise tracing one feature from the screen to the service it uses. If you commission an app, ask who watches notices like this and follows through on the update.