Google AI Edge Gallery
Try Gemma 4 and other LLMs fully on-device — no cloud, no account, no API key
Google AI Edge Gallery is an open-source Android app and showcase repository that lets users run on-device ML and generative AI models — including Gemma 4 E2B and E4B — directly on their phone with no cloud dependency, no API key, and no data leaving the device. It launched on April 8, 2026, alongside LiteRT-LM, and serves as both a consumer demo and a developer reference implementation for building private, offline-first AI applications. The Gallery app installs from the Play Store and provides a clean interface to select models, run them locally, and explore use cases across text generation, image analysis, and on-device tool use. For developers, the open-source codebase is a working example of how to integrate LiteRT-LM into a production Android app — showing everything from model download and caching to GPU/NPU acceleration and streamed token output. The broader significance is Google's public commitment to on-device AI as a first-class product surface. By shipping a polished Play Store app rather than just a GitHub repo, Google is telling device manufacturers, enterprise security teams, and privacy-conscious developers that running Gemma locally is a supported, maintained path — not a research experiment. With Gemma 4 E4B fitting under 3GB and achieving competitive benchmark performance, the 'local AI' narrative just got a major mainstream distribution vehicle.
Panel Reviews
“”
“”
“”
“”
Community Sentiment
“”
“”
“”