🤖

Enabling Gemini Nano in Chrome

AI & Autonomous
Listen to this page
⚡ Quick Copy Helper

Web security models restrict pages from launching chrome:// pages automatically. Use the buttons below to copy each address, then paste them directly into your browser's address bar.

1. Enable Device Model Execution Flag
chrome://flags/#optimization-guide-on-device-model
2. Enable Prompt API Flag
chrome://flags/#prompt-api-for-gemini-nano
3. Go to Components (Download Component)
chrome://components/

Enabling Gemini Nano in Chrome

Gemini Nano is built directly into the Google Chrome browser and does not require a manual download. It executes locally on your device for fast, private, offline-capable intelligence.

Once activated, Chrome will automatically download a ~4 GB language model in the background.

Prerequisites

  • Chrome Version: Ensure you are running Chrome Canary (version 128+) or the latest stable version of Google Chrome.
  • Hardware Requirements: A machine with sufficient RAM (typically 8 GB or more) and a compatible GPU/CPU.

Step-by-Step Enablement Guide

1. Enable Experimental Flags

To enable Chrome's built-in AI APIs, open a new browser tab and configure the following flags:

1. Optimization Guide On-Device Model:

- Navigate to: chrome://flags/#optimization-guide-on-device-model

- Set it to: Enabled BypassPrefRequirement (or Enabled BypassPerfRequirement depending on your version). This enables the optimization guide to execute models on device across Mac, Windows, Linux, and ChromeOS.

2. Prompt API for Gemini Nano:

- Navigate to: chrome://flags/#prompt-api-for-gemini-nano

- Set it to: Enabled. This enables the exploratory Prompt API, allowing Phasoric to send natural language instructions to the built-in Gemini Nano model.

3. Optional Companion APIs (for extended capabilities):

- Prompt API with Multimodal Input: chrome://flags/#prompt-api-for-gemini-nano-multimodal-input → Enabled

- Writer API: chrome://flags/#writer-api-for-gemini-nano → Enabled

- Rewriter API: chrome://flags/#rewriter-api-for-gemini-nano → Enabled

- Proofreader API: chrome://flags/#proofreader-api → Enabled

- Classifier API: chrome://flags/#classifier-api-for-tiny-model → Enabled

- Summarizer API: chrome://flags/#summarizer-api-for-gemini-nano → Enabled

- Summarizer Performance Preference: chrome://flags/#summarizer-api-performance-preference → Enabled

2. Restart your Browser

After changing the flags, click the blue Relaunch button at the bottom of the page to apply the changes.

3. Trigger the Model Download

Once restarted, you must trigger Chrome to download the underlying Gemini Nano model:

1. Navigate to: chrome://components/

2. Scroll down or search to find the component named Optimization Guide On Device Model.

3. Click the Check for update button.

4. The status should change to Downloading and eventually Up-to-date (once the ~4 GB model download completes in the background).

Verification

After the download finishes, return to your Phasoric settings page and check the status. Browser intelligence should now show as available and ready for use!