Google Launches Local AI Model App for Offline Use
Last week, Google discreetly launched an application that enables users to run various openly accessible AI models from Hugging Face's development platform directly on their mobile devices.
Named Google AI Edge Gallery, this application is currently available for Android and will soon arrive on iOS. It empowers users to discover, download, and execute compatible models capable of generating images, answering questions, writing and editing code, among other tasks. These models operate entirely offline, utilizing the processors of supported smartphones without requiring an internet connection.
While cloud-based AI models typically offer greater power than local versions, they come with certain limitations. Some users may hesitate to transmit personal or sensitive information to remote data centers, or prefer having models accessible without relying on Wi-Fi or cellular networks.

The welcome interface of Google AI Edge Gallery for Android.Image Credits:Google Google describes AI Edge Gallery as an "experimental Alpha release," available for download from GitHub by following specific instructions. The main interface presents shortcuts to AI functions like "Ask Image" and "AI Chat." Selecting any function displays a list of suitable models for that task, including Google's Gemma 3n.
The application also features a "Prompt Lab" where users can initiate single-step tasks using various models, such as text summarization and rewriting. This lab includes multiple task templates and adjustable settings to customize model behavior.
Google notes that performance may vary depending on device capabilities. While modern devices with advanced hardware naturally process models faster, the model size significantly impacts performance. Larger models require more time to complete tasks—like answering image-related questions—compared to their smaller counterparts.
Google encourages developer community members to share feedback about their AI Edge Gallery experience. The application operates under an Apache 2.0 license, permitting unrestricted use in most commercial and non-commercial contexts.
Register Now Through June 4 for TechCrunch Sessions: AI
Secure $300 savings on TC Sessions: AI tickets—plus 50% off a second pass. Gain insights from industry leaders at OpenAI, Anthropic, Khosla Ventures, and others during a comprehensive day of expert analysis, practical workshops, and valuable networking opportunities. These special rates expire when the event begins on June 5.
Exhibit at TechCrunch Sessions: AI
Reserve your exhibition space at TC Sessions: AI and present your innovations to 1,200+ decision-makers without premium costs. This offer remains available through May 9 or until booth capacity is reached.
Berkeley, CA|June 5REGISTER NOW
Related article
Google IO 2026 unveils voice interaction with Gmail inbox
Google continues to integrate AI into your inbox. At the IO 2026 developer conference on Tuesday, the company expanded its Gmail "AI Inbox" feature with conversational AI, allowing users to ask questions about their inbox content rather than relying
WordPress.com now allows AI agents to write and publish posts, plus more
WordPress.com, the popular web hosting and publishing platform, is now embracing AI agents—a move that could reshape the look and feel of the web. The company announced Friday that it will allow AI agents to draft, edit, and publish content on custom
Kakao Mobility outlines Level 4 autonomous driving roadmap for physical AI
Kakao Mobility is planning to develop Level 4 autonomous driving technologies internally as part of its physical AI strategy.
At the 2026 World IT Show conference in Seoul's COEX, Kim Jin-kyu — vice president and head of Kakao Mobility's Physical AI
Related Special Topic Recommendations
Comments (0)
0/500
Last week, Google discreetly launched an application that enables users to run various openly accessible AI models from Hugging Face's development platform directly on their mobile devices.
Named Google AI Edge Gallery, this application is currently available for Android and will soon arrive on iOS. It empowers users to discover, download, and execute compatible models capable of generating images, answering questions, writing and editing code, among other tasks. These models operate entirely offline, utilizing the processors of supported smartphones without requiring an internet connection.
While cloud-based AI models typically offer greater power than local versions, they come with certain limitations. Some users may hesitate to transmit personal or sensitive information to remote data centers, or prefer having models accessible without relying on Wi-Fi or cellular networks.

Google describes AI Edge Gallery as an "experimental Alpha release," available for download from GitHub by following specific instructions. The main interface presents shortcuts to AI functions like "Ask Image" and "AI Chat." Selecting any function displays a list of suitable models for that task, including Google's Gemma 3n.
The application also features a "Prompt Lab" where users can initiate single-step tasks using various models, such as text summarization and rewriting. This lab includes multiple task templates and adjustable settings to customize model behavior.
Google notes that performance may vary depending on device capabilities. While modern devices with advanced hardware naturally process models faster, the model size significantly impacts performance. Larger models require more time to complete tasks—like answering image-related questions—compared to their smaller counterparts.
Google encourages developer community members to share feedback about their AI Edge Gallery experience. The application operates under an Apache 2.0 license, permitting unrestricted use in most commercial and non-commercial contexts.
Register Now Through June 4 for TechCrunch Sessions: AI
Secure $300 savings on TC Sessions: AI tickets—plus 50% off a second pass. Gain insights from industry leaders at OpenAI, Anthropic, Khosla Ventures, and others during a comprehensive day of expert analysis, practical workshops, and valuable networking opportunities. These special rates expire when the event begins on June 5.
Exhibit at TechCrunch Sessions: AI
Reserve your exhibition space at TC Sessions: AI and present your innovations to 1,200+ decision-makers without premium costs. This offer remains available through May 9 or until booth capacity is reached.
Berkeley, CA|June 5REGISTER NOW
Google IO 2026 unveils voice interaction with Gmail inbox
Google continues to integrate AI into your inbox. At the IO 2026 developer conference on Tuesday, the company expanded its Gmail "AI Inbox" feature with conversational AI, allowing users to ask questions about their inbox content rather than relying
WordPress.com now allows AI agents to write and publish posts, plus more
WordPress.com, the popular web hosting and publishing platform, is now embracing AI agents—a move that could reshape the look and feel of the web. The company announced Friday that it will allow AI agents to draft, edit, and publish content on custom





Home






