Last week, Google quietly released an app called Google AI Edge Gallery that lets users run a range of openly available AI models from the AI development platform Hugging Face right on their phones. The app is currently available for Android and is expected to come to iOS soon.
Run AI Models Locally Without Internet
This app allows users to find, download, and run compatible AI models that perform tasks such as generating images, answering questions, writing and editing code, and more — all offline. These models run directly on supported phones’ processors, so they don’t require an internet connection to operate.
Cloud-based AI models are often more powerful, but they come with downsides. Some users worry about sending personal or sensitive data to remote servers, and others prefer to use AI tools without needing to find Wi-Fi or cellular access.
Upon opening the app, the home screen offers shortcuts to AI tasks like “Ask Image” and “AI Chat.” Selecting a capability presents a list of models suited for that purpose, including Google’s own Gemma 3n.
Prompt Lab for Custom AI Tasks
Google AI Edge Gallery also includes a “Prompt Lab,” where users can initiate “single-turn” AI tasks such as summarizing or rewriting text. The Prompt Lab comes with pre-set templates and adjustable settings that let users fine-tune how the models respond.
Google cautions that performance will vary depending on your device’s hardware. More powerful phones will run the models faster. Additionally, larger models take longer to complete tasks — for example, answering a question about an image may take more time with bigger models than with smaller ones.
Google is inviting developers to try the app and provide feedback on their experience. Currently labeled an “experimental Alpha release,” the app is available for download on GitHub with detailed instructions for installation.
The app is released under the Apache 2.0 license, meaning it can be freely used in most contexts, including commercial purposes, without restrictions.
Author’s Opinion
The ability to run AI models locally on devices like smartphones marks a significant step toward enhancing user privacy and control. Relying solely on cloud-based AI means constantly sending data to remote servers, which raises concerns about security and data misuse. With apps like Google AI Edge Gallery, users gain the freedom to explore powerful AI tools without compromising their privacy or connectivity. This development could encourage wider adoption of AI by making it more accessible and trustworthy.
