Google Unveils AI Edge Gallery App for Running AI Models Locally

Google has launched the AI Edge Gallery, a groundbreaking Android application designed to enable local execution of artificial intelligence (AI) models. Unveiled as an experimental alpha release, this app empowers users to download, run, and interact with AI models such as large language models (LLMs) directly on their smartphones—without requiring an internet connection.

App Availability and Installation Details

Currently available on GitHub, the AI Edge Gallery is released under the open-source Apache 2.0 license, allowing both commercial and academic use. Users can download the Android application package (APK) file, which is approximately 115MB in size. The app comes with a detailed installation guide to help users set it up on compatible devices.

Although installing APKs from outside the Google Play Store can carry potential security risks, initial testing by Gadgets 360 staff indicated that the app is free from malware or viruses. Nonetheless, users are advised to take standard precautions when sideloading APK files.

Core Features of the Google AI Edge Gallery App

The app currently includes three key AI-powered functionalities:

1. Ask Image: This feature allows users to upload an image and ask questions about it. Once an appropriate model is run locally, the AI provides analysis and insights based on the image content.

2. AI Chat: Users can engage in conversations with the on-device LLMs. However, since the models run locally, the responses may not reflect real-time or updated information.

3. Prompt Lab: Designed for more advanced users, this feature enables tone-based rewriting, text summarization, creative generation, and code snippet creation. Users can fine-tune model behavior by adjusting settings like token limits, temperature, and accelerator preferences. Benchmark metrics are also available for reference.

Device Compatibility and Model Availability

The app supports a variety of AI models, but performance and compatibility largely depend on the device's hardware capabilities. Devices equipped with AI-focused chipsets will experience better efficiency and broader functionality. In addition to downloading models from within the app, users can also import models that they’ve previously acquired.

Privacy and Offline Usage

One of the most notable advantages of the Google AI Edge Gallery app is that all operations are conducted locally. This ensures that user data never leaves the device, offering an added layer of privacy and eliminating the need for constant internet access. The offline-first approach is particularly beneficial for users concerned about data security and latency.

Planned Expansion: iOS Version Coming Soon

Google has confirmed that an iOS version of the AI Edge Gallery app is in development and will be available soon. This upcoming release will expand the reach of the app to Apple’s ecosystem, further democratizing access to on-device AI functionalities.

Future Enhancements and Expectations

As the app is still in its alpha phase, more features and enhancements are expected over time. Future updates may include support for additional models, improved processing capabilities, and a broader suite of AI tools. Google’s move reflects the broader industry trend of bringing powerful AI tools to the edge—making them more accessible, private, and responsive.

Conclusion: Making AI Personal and Private

The AI Edge Gallery app by Google is a significant milestone in the advancement of localized AI technology. By enabling offline access to LLMs and other models, Google is pushing the boundaries of what’s possible on mobile devices. As the platform evolves, it promises to become a critical tool for developers, researchers, and everyday users who want to tap into the power of artificial intelligence—without relying on the cloud.