Now that the Generative AI for Android is deprecated, I can’t find any supported ways to use Gemini API in Android.
The README states to use Firebase SDK, Which only supports Vortex AI. I guess it’s great for large apps, but for my use cases - Gemini API is more than enough, and I do not want to use Vortex.
I think it’s kind of ridiculous that they assume everything in Android is production when other languages (e.g. Python, TypeScript, Go) has perfectly fine SDK made for this purpose.
Am I missing something?
3 Likes
Hey @azu0609 . Welcome to the community .
You’re not missing anything obvious in your assessment. The current official guidance for integrating the Gemini API into Android apps does indeed point towards the Vertex AI integration within the Firebase ecosystem.
If your use case aligns with the capabilities of Gemini Nano (currently focused on text-based tasks) and you’re targeting compatible devices, the Google AI Edge SDK offers an on-device solution. Keep in mind this is currently experimental.For more information you can refer this Gemini Nano with the Google AI Edge SDK | Android Developers
While I don’t have immediate news about a standalone, lightweight Gemini API SDK for Android, we will get back to you with more information as soon as possible.
Thank you for response.
Gemini Nano would probably be enough, but I don’t own any pixel devices so I can’t even test them in the first place…
Hope you guys come up with new SDK for Android soon.
I understand the limitation with testing Gemini Nano without a Pixel device – that’s definitely a barrier. Your feedback is noted and has been escalated to the senior development team. We’re always striving to improve accessibility for all developers.Thank you for bringing attention to this issue.