In a low-profile move, Google recently rolled out a new app that enables users to download and run AI models directly on their smartphones — without needing an internet connection.
The app, called Google AI Edge Gallery, is currently available for Android, with an iOS version in the works. It gives users access to a variety of AI models hosted on Hugging Face, allowing them to perform tasks like generating images, answering questions, editing code, and more — all offline. The models tap into the phone’s onboard processors, bypassing the need for cloud computing.
While cloud-based AI tends to offer higher performance, running models locally has its advantages. It provides more privacy by keeping data on the device and ensures functionality even without Wi-Fi or mobile data.
Billed as an “experimental Alpha release,” the app can be downloaded via GitHub, along with setup instructions. Its interface features shortcuts to tools such as “Ask Image” and “AI Chat.” Selecting one opens a list of compatible models — including Google’s own Gemma 3n — tailored for that task.
A notable feature is the “Prompt Lab,” where users can launch single-turn AI actions like rewriting or summarizing text. It includes task templates and adjustable settings to customize model behavior.
However, Google notes that performance will vary. More advanced devices will naturally handle models faster, but model size is also a factor — larger models may take longer to complete tasks such as interpreting images or generating responses.
The company is encouraging developers to test the app and share feedback. Released under the Apache 2.0 license, the Google AI Edge Gallery can be freely used in both personal and commercial projects.
We have helped 20+ companies in industries like Finance, Transportation, Health, Tourism, Events, Education, Sports.