Recent updates from Google on the Gemini Nano API via ML kit are important for developers who aim to integrate the generated AI features into Android applications. The new Genai API allows for inference on the device. This means that all operations, including inputs, processes, and output, are performed locally on the device. This design not only enhances user privacy, but also eliminates the costs associated with cloud processing.
Developers can use the Genai API to take advantage of the following features:
- summary: You can summarise articles in a concise format with up to 3,000 English words.
- Calibration: This feature improves text for grammar accuracy and consistency in style.
- Rewrite: Text can be rewritten in a variety of tones and styles, such as formal, casual, and friendly.
- Image description: Generates a brief description of the image and improves accessibility.
For practical implementations, developers can integrate the summary API using the following Java snippet:
val articleToSummarize = ...
val summarizerOptions = SummarizerOptions.builder(context)
.setInputType(InputType.ARTICLE)
.setOutputType(OutputType.ONE_BULLET)
.setLanguage(Language.ENGLISH)
.build()
val summarizer = Summarization.getClient(summarizerOptions)
val summarizationRequest = SummarizationRequest.builder(text).build()
summarizer.runInference(summarizationRequest) { newText ->
// Show new text in UI
}
For more information on how to get started, see the official ML Kit Genai APIS documentation.
Device compatibility and performance
The ML Kit Genai API is optimized for a variety of Android devices, especially Android devices powered by MediaTek Dimenity, Qualcomm Snapdragon, and Google Tensor Platform. Supported devices include the Pixel 9 series, Samsung Galaxy S25, and more. This will broaden the application of the Genai API beyond GEAL's hardware alone, and encourage wider adoption across the Android ecosystem.
The Genai API is designed for efficiency in running on-device without the need for an internet connection. This ensures a smooth user experience and quick response time, as it does not rely on external servers.
Developers can expect to see performance benchmarks for tasks such as summaries, calibrations, and rewrites that significantly improve over the basic Gemini Nano model. The API is built to deliver high quality results immediately, minimizing the need for extensive rapid engineering.
Privacy and security
One of the outstanding features of the Genai API is the enhancement of privacy they offer. By processing data on your device locally, developers can ensure that sensitive information remains secure to users. This is particularly relevant for applications that require processing personal data, such as email and text communications.
The ML kit safety protocols include various protective layers to prevent misuse. Includes classifiers and evaluations to ensure the integrity of the data being processed.
Conclusion
The introduction of the ML Kit Genai API allows developers to easily integrate powerful AI capabilities into their applications while maintaining user privacy. This important advancement in AI technology on devices is consistent with the growing demand for secure and efficient mobile applications.
To find out how to power your application with a passwordless authentication solution, visit Mojoauth. Our services provide a smooth and secure login experience for web and mobile applications, leaving your user data protected while simplifying access.
***This is Mojoauth's Security Blogger Network Syndicate Blog. You will be redirected to the passwordless created by DeveshPatel. https://mojoauth.com/blog/google-unveils-gemini-nano-for-on-device-ai-in-android-apps/
