The application, called Google AI Edge Gallery, is now available on the Android platform and will be released to iOS soon.
Google AI Edge Gallery allows users to access open-source AI models from the AI hugging Face development platform, performing many tasks such as creating images, answering questions, writing and editing programming code, and many other functions.
These models work completely online, taking advantage of the processing power of supported phones, helping users not need to rely on Wi-Fi connection or mobile data.
A prominent advantage of running local AI is privacy. Users can feel more secure when personal data is not sent to remote data centers. While AI models on cloud platforms are often stronger, on-site solutions like Google AI Edge Gallery provide more flexibility and control for users.
The application is in the testing phase with the Alpha label and can be downloaded from GitHub under specific instructions. The main interface of the application provides shortcomings to AI tasks such as AI Chat or Ask Image. When a user selects a function, the application will display a list of models that match that task, such as Gemma 3n developed by Google itself.
Another notable feature is the Prompt Lab feature, where users can experiment with individual tasks such as summarizing or re-writing text. Prompt Lab provides many available task templates and settings to customize the model's operation.
Google notes that the performance of the app can vary depending on the device. Modern phones with powerful hardware will run faster models. However, model size is also an important factor: large models may take more time to complete tasks, such as answering questions related to images.
The application is released under the Apache 2.0 license, allowing free use in many context, including commercial. Google also calls on the developer community to contribute ideas to improve the AI Edge Gallery experience.
