Google AI Edge Gallery
A mobile app for running open-source generative AI models fully on-device, for exploring, trying and evaluating on-device LLMs and other generative AI features offline.
Google AI Edge Gallery is a mobile app for running open-source large language models and other generative AI features directly on a device, with all processing offline and local. It supports the Gemma model family, including the Gemma 4 release, and installs from Google Play, the App Store, or a downloadable APK and macOS build.
Core features include Agent Skills, which augment a loaded model with tools such as Wikipedia lookups, interactive maps and visual summary cards; a chat mode with an optional Thinking Mode that shows a supported model's step-by-step reasoning; Ask Image, for identifying objects or answering questions about a photo or camera view using multimodal models; Audio Scribe, for on-device transcription and translation of voice recordings; and a Prompt Lab for testing prompts with direct control over parameters such as temperature and top-k. Mobile Actions and Tiny Garden use a fine-tuned small model for offline device control and a natural-language gardening mini-game.
It also includes model management and benchmarking tools for comparing on-device models. It is written mainly in Kotlin and licensed Apache-2.0.