Templates

On-device assistant

Quick explainer2 min readIntermediate

An on-device assistant runs its AI model on your own phone, laptop or browser instead of a cloud server, so it can work offline and keep your data local.

In plain words

A small model ships with the device or is downloaded once, and apps call it through system APIs. On Android, Gemini Nano runs inside a system service called AICore and supports tasks such as summarising and proofreading. Chrome offers built-in AI APIs where the browser supplies and manages the model.

A quick example

A messaging app uses the phone's on-device model to summarise a long group chat into bullet points. Because the model runs locally, the chat never leaves the phone and the feature works without a network connection.

Where you'll meet it

  • Private features on phones, like summaries and smart replies.
  • Web apps that translate or summarise in the browser.

Related

Sources

  1. DocsGemini Nano, Android Developers · read 8 Oct 2026
  2. DocsBuilt-in AI, Chrome for Developers · read 29 Sept 2026

Checked by WhatIsAI against the sources above · 8 Oct 2026. Spotted an error? Report it.