Google’s AI Edge Gallery app launches on Android, iOS version coming soon

Google’s new AI Edge Gallery app lets users run powerful AI models offline on Android devices, enhancing privacy and speed by eliminating internet dependency. Featuring the compact Gemma 3 1B model and open-source access, it promises broader AI accessibility with iOS support on the horizon.

Sources:
Android PoliceIndiatoday
Updated 1h ago
Tab background
Sources: Android PoliceIndiatoday
Google has introduced AI Edge Gallery, an experimental app currently available on Android that enables users to run AI models directly on their devices without requiring Wi-Fi or cellular connections.

At the heart of the app is the compact language model Gemma 3 1B, which weighs just 529MB and can process up to 2,585 tokens per second, allowing for rapid text generation and smooth interactions.

The app integrates open-source AI models from Hugging Face and leverages Google’s AI Edge platform technologies like TensorFlow Lite and MediaPipe to optimize performance across a wide range of devices, including those with modest hardware.

Running AI locally offers significant advantages: faster responses due to no server lag and enhanced privacy since data never leaves the device, reducing risks of interception or misuse.

Although still in an "experimental Alpha release," AI Edge Gallery is fully open source under the Apache 2.0 license, allowing developers and companies to freely use, modify, or integrate it into commercial products.

Google’s move positions it as a leader in offline AI, making powerful models accessible from users’ pockets without internet dependency. The app is not yet on the Play Store but is available via GitHub with an official user guide. An iOS version is expected soon, expanding accessibility.

"Despite that, the release of AI Edge Gallery positions Google as a leader in making offline AI not only possible but also practical," highlighting a step toward more personal, private, and always-ready AI experiences.
Sources: Android Police
Google has launched AI Edge Gallery, an experimental app for Android that runs AI models locally without internet, enhancing speed and privacy. The app features Google’s Gemma 3n and open-source models from Hugging Face, with an iOS version expected soon.
Section 1 background
With the ability to run powerful models right from your pocket, it’s a step toward making AI more personal, private, and always ready — no Wi-Fi required.
Google spokesperson
Indiatoday
Key Facts
  • Google has launched AI Edge Gallery, an experimental app that allows users to run AI models locally on their devices without needing an internet connection.Android Police
  • The app supports open-source AI models from Hugging Face and includes Google’s own Gemma 3 1B language model.Android PoliceIndiatoday
  • The Gemma 3 1B model is compact at 529MB and can process 2,585 tokens per second, enabling fast text generation and seamless interactions.Indiatoday
  • AI Edge Gallery is built on Google’s AI Edge platform using TensorFlow Lite and MediaPipe to optimize performance across a wide range of devices, including those with modest hardware.Indiatoday
  • The app is released as an experimental Alpha version and is fully open source under the Apache 2.0 license, allowing developers and companies to use, modify, or integrate it into commercial products.Indiatoday
  • AI Edge Gallery is currently available on Android via GitHub and is not yet on the Play Store; an iOS version is coming soon.Android Police
Key Stats at a Glance
Size of Gemma 3 1B language model
529MB
Indiatoday
Processing speed of Gemma 3 1B model
2,585 tokens per second
Indiatoday
Article not found
⌂ CuriousCats.ai

Article

‹

Source Citations