The rapid expansion of the generative AI has changed the way Google and other technology giants design products, but most of the AI ​​features you have used work on remote servers with a ton of treatment power. Your phone has much less power, but Google seems ready to give developers new important mobile AI tools. At E / S next week, Google will probably announce a new API set to allow developers to take advantage of the capacities of Gemini Nano for an AI on the device.
Google discreetly published the documentation on the major AI features for developers. According to Android authorityAn update of the SDK of the ML kit will add the management of the API for generative functionalities on AI devices via Gemini Nano. It is built on the AI ​​nucleus, similar to EDGE experimental SDK, but it connects to an existing model with a set of predefined features which should be easy to implement for developers.
Google says that ML Kit's Genai APIs will allow applications to make a summary, a rereading, a rewriting and an image description without sending data to the cloud. However, Gemini Nano does not have as much power as the cloud-based version, so expect certain limitations. For example, Google notes that summaries can only have a maximum of three chips and that image descriptions will only be available in English. The quality of the outputs could also vary depending on the version of Gemini Nano on a phone. The standard version (Gemini Nano XS) measures approximately 100 MB, but Gemini Nano XXS, as seen on the 9a pixel, is a quarter of the size. It is only in text and has a much smaller context window.
Not all versions of Gemini Nano are also created.
Credit: Ryan Whitwam
This decision is good for Android in general because ML KIT works on devices outside the Google Pixel line. While the Pixel devices largely use Gemini Nano, several other phones are already designed to execute this model, including the OnePlus 13, Samsung Galaxy S25 and Xiaomi 15. While more and more phones add a management of the Google IA model, developers will be able to target these devices with generative AI features.
