Google has quietly released the 'Google AI Edge Gallery' app, which lets you download AI models to your smartphone and run them locally.
Google has released the app ' Google AI Edge Gallery ' on May 21, 2025, which allows users to run AI models published on the AI development platform Hugging Face from their smartphones. At the time of writing, the Android version of Google AI Edge Gallery has been released, and the iOS version is scheduled to be released soon.
GitHub - google-ai-edge/gallery: A gallery that showcases on-device ML/GenAI use cases and allows people to try and use models locally.
Google quietly released an app that lets you download and run AI models locally | TechCrunch
https://techcrunch.com/2025/05/31/google-quietly-released-an-app-that-lets-you-download-and-run-ai-models-locally/
Google AI Edge is a group of technologies that enable efficient execution of AI models on edge devices such as smartphones and IoT devices. Traditionally, AI models were typically executed on high-performance servers, but Google says that by using Google AI Edge, AI processing can be performed directly on the device even in environments without an Internet connection.
Google AI Edge Gallery is an app released for smartphones that allows users to search for AI models that can generate images, answer questions, code, etc., and download and run compatible models. The models run locally on the smartphone's processor without requiring an internet connection.
The Android version released is an 'experimental alpha version,' and the APK file is distributed on the GitHub page . For specific installation and configuration instructions, please read the following article.
Introducing the Google-made open source AI model 'Gemma 3n' that runs locally on smartphones, here's how to use it on your smartphone right now - GIGAZINE
The home screen shows shortcuts to AI tasks and features, like 'Ask Image' and 'AI Chat.' Tapping on a feature brings up a list of models suitable for that task, such as Google's Gemma 3n.
Below is 'Ask Image' asking Gemma 3n to solve a math problem.
You can also talk to Gemma 3 using 'AI Chat.'
The Google AI Edge Gallery also features 'Prompt Lab,' which allows users to initiate 'single-turn' tasks with the model, such as summarizing or rewriting text. Prompt Lab offers multiple task templates and settings to fine-tune the model's behavior.
Naturally, Google warns that 'performance results may vary' and that models run faster, as expected, on modern devices with more powerful hardware, but model size also matters: the larger the model, the longer it will take to complete the task.
Google invites members of the developer community to provide feedback on their experiences with the Google AI Edge Gallery, which is licensed under the Apache 2.0 license and can be used without restrictions in most situations, both commercial and non-commercial.
Related Posts:
