Image 1 of 10
Build Smarter Mobile & Web Apps: Integrate Gemini Using Firebase AI Logic
Google for Developers Day: İstanbul
The session opens with the AI applications most of us already use every day, and what an AI feature looks like from the developer side of them. Firebase AI Logic comes next. It gives client applications a direct path to Gemini models without running a backend, with SDKs for Android, iOS, Web, Flutter, and Unity.
We then lay out the structure of the app we build in the session, a recipe generator where a list of ingredients goes in and a recipe with a generated image comes out. The text call and the image call stay separate, and you will see what that separation buys you in structured output and quality control. The integration is then built step by step.
Best practices close the technical part. App Check verifies that a request reaching Gemini came from your real app, and we cover what changes when enforcement becomes a requirement for Firebase AI Logic. The last minutes cover Prompt Templates and what changes when the model name lives outside the build. The session ends with the finished app running live.














