Google is set to broaden the horizons of its Gemini Nano, an innovative on-device AI model, through the introduction of new ML Kit GenAI APIs. This expansion, anticipated to be unveiled at the upcoming I/O 2025 event, promises to empower developers with the ability to seamlessly integrate features such as text summarization, proofreading, rewriting, and image description generation into their applications.
Enhancing User Experience with On-Device AI
Generative AI technology is reshaping the landscape of online communication and content creation. Users frequently turn to AI chatbots like Google Gemini for assistance with tasks ranging from article summarization to email proofreading. However, concerns regarding privacy—especially when handling sensitive information—have led to a cautious approach among some users. In response, Google has developed Gemini Nano, a compact and optimized version of its AI model that operates directly on devices rather than relying on cloud servers. This shift not only enhances privacy but also marks a significant step towards making advanced AI capabilities more accessible.
Recently, Google revealed a session titled “Gemini Nano on Android: Building with on-device gen AI” in its I/O 2025 session list. The session aims to introduce a new suite of generative AI APIs that leverage the capabilities of Gemini Nano. These APIs are designed to simplify the implementation of various use cases, including text summarization, proofreading, rewriting, and image description generation.
In October, Google initiated experimental access to Gemini Nano via the AI Edge SDK, allowing developers to explore text-to-text prompts specifically on the Pixel 9 series. While the AI Edge SDK facilitated features such as rephrasing and smart replies, it notably lacked support for generating image descriptions—an essential feature highlighted for the forthcoming I/O session. This suggests that the upcoming generative AI APIs may either represent a new offering or an enhanced iteration of the existing AI Edge SDK.
Earlier this week, Google quietly released documentation for the ML Kit’s new GenAI APIs. This SDK enables developers to harness machine learning capabilities within their applications without needing a deep understanding of the underlying models. The GenAI APIs allow developers to utilize Gemini Nano’s power for common tasks through a user-friendly interface, ensuring that data processing remains local to enhance privacy and performance.
Features of ML Kit’s GenAI APIs
The ML Kit’s GenAI APIs provide a range of features that align with the functionalities expected from the upcoming I/O session:
- Summarization: Generate concise summaries of articles or chat conversations as bulleted lists, producing up to three bullet points in languages including English, Japanese, and Korean.
- Proofreading: Refine short content by correcting grammar and spelling errors, available in multiple languages such as English, Japanese, German, French, Italian, Spanish, and Korean.
- Rewrite: Transform chat messages into different tones or styles, offering options like Elaborate, Emojify, Shorten, Friendly, Professional, and Rephrase in several languages.
- Image Description: Generate brief descriptions for images, currently supported in English.
Unlike the existing AI Edge SDK, which is classified as “experimental access,” the ML Kit’s GenAI APIs will be available in “beta.” This designation indicates a potential for production use, allowing developers to release applications that incorporate these new capabilities. Furthermore, while the AI Edge SDK is restricted to text input, the GenAI APIs will support image input, enabling the much-anticipated image description feature.
Perhaps the most significant distinction lies in device compatibility. The AI Edge SDK is limited to the Google Pixel 9 series, whereas the ML Kit’s GenAI APIs will be accessible on any Android device that supports the multimodal Gemini Nano model. This broadens the scope to include a variety of devices, such as the HONOR Magic 7, Motorola Razr 60 Ultra, OnePlus 13, Samsung Galaxy S25, Xiaomi 15, and more.
For developers eager to explore the capabilities of Gemini Nano within their applications, the public documentation for the ML Kit GenAI APIs is now available, paving the way for innovative integrations in the near future.
Got a tip? Talk to us!
Email our staff at news@androidauthority.com. You can stay anonymous or get credit for the info, it’s your choice.
