Google Quietly Launches AI Edge Gallery App for Offline AI Model Execution

Google has quietly released Google AI Edge Gallery, an Android app that enables users to download and run various openly available AI models locally on their smartphones. An iOS version is expected to follow soon. This app allows users to execute AI tasks without an internet connection, leveraging the processing power of their devices.

The AI Edge Gallery app sources its models from the Hugging Face AI development platform. Users can find, download, and run models capable of generating images, answering questions, writing and editing code, and performing other AI-driven tasks directly on their devices.

While cloud-based AI models often offer more power, running models locally provides distinct advantages. Users concerned about data privacy can avoid sending sensitive information to remote servers. Furthermore, local execution ensures model availability even without Wi-Fi or cellular connectivity.

Google describes AI Edge Gallery as an “experimental Alpha release.” The app is available for download from GitHub, with installation instructions provided. Once installed, users can access AI capabilities such as “Ask Image” and “AI Chat.” The app supports models like Google’s Gemma 3n, optimized for mobile devices.

The AI Edge Gallery includes a “Prompt Lab” for single-turn tasks like text summarization and rewriting. This feature provides task templates and configurable settings to fine-tune model behavior.

Google advises that performance may vary depending on device hardware and model size. Modern, more powerful devices will predictably run models faster. Smaller models will generally complete tasks more quickly than larger ones.

Google is actively soliciting feedback from the developer community to improve the AI Edge Gallery experience. The app is released under the Apache 2.0 license, allowing for unrestricted use in both commercial and non-commercial contexts.