Creating images, answering questions, writing and editing code can be done directly on the phone without the Internet (photo: Google).
Google recently released AI Edge Gallery , an experimental application that allows deploying and using artificial intelligence (AI) models directly on mobile devices, without requiring a network connection.
The app serves as a library of AI models, mostly sourced from the open-source Hugging Face platform. Users can download and run these models right on their devices, with some common tasks like answering questions via images or chatting with AI.
Unlike cloud-based AI systems, AI Edge Gallery processes data locally, minimizing the transmission of information to the outside, thereby contributing to limiting the risk of personal data leakage.
One notable feature is Prompt Lab , where users can quickly initiate AI tasks like text summarization or content rewriting. Prompt Lab offers a variety of setups and customizations to tailor the model's behavior to specific goals.
The app also supports AI models of varying sizes and complexity. Processing performance depends on the device's hardware configuration – newer phones with powerful processors will produce faster results, while older devices may experience lag or feature limitations.
Google has now released the AI Edge Gallery on GitHub under the Apache 2.0 license. Users can download it and use it for a variety of purposes, including commercial development.
The project is in open beta, encouraging the developer community to contribute ideas.
Source: https://dantri.com.vn/cong-nghe/nguoi-dung-dien-thoai-android-co-the-su-dung-ai-ma-khong-can-internet-20250601003306528.htm
Comment (0)