Offline AI just went mainstream with Google’s new tool
Google released the Google AI Edge Gallery app last week, enabling users to download and run AI models from Hugging Face on their phones. The Android app, with an iOS version forthcoming, facilitates offline AI task execution.
The Google AI Edge Gallery, labeled an “experimental Alpha release,” is accessible for download via GitHub, accompanied by installation instructions. This app allows users to access various AI capabilities such as “Ask Image” and “AI Chat”. Selection of a capability displays a list of appropriate models, including Google’s Gemma 3n.
The application allows users to locate, download, and run compatible AI models capable of generating images, answering questions, and writing/editing code. These models operate offline, utilizing the processing power of supported phones.
The AI Edge Gallery also features a “Prompt Lab” for initiating single-turn tasks such as text summarization and rewriting. This lab provides task templates and adjustable settings for model customization.
Can Google’s tiny Gemma 3n AI really run smoothly on any device?
Google cautions that performance may vary. Devices with more powerful hardware are expected to run models faster. Model size influences task completion time; larger models require more processing time.
Google is soliciting feedback from the developer community regarding the Google AI Edge Gallery experience. The app operates under an Apache 2.0 license, permitting usage in most contexts, including commercial applications, without limitations.