Skip to main content
The AI Toolbar is VAssist’s context-aware interface that appears when you select text or hover over images on any website. It provides instant access to powerful AI features powered by Chrome’s built-in Gemini Nano model.
AI Toolbar in action

Text Selection Features

Select any text on any webpage to reveal the AI toolbar with these features:

Summarize

Turn walls of text into digestible chunks using Chrome’s Summarizer API. Get three types of summaries:
  • Headline - One punchy line that captures the essence
  • Key Points - Bullet list of main ideas
  • Teaser - Short preview that hooks you in
// Chrome's built-in Summarizer API
const summarizer = await self.aiSummarizer.create({
  type: 'tldr',
  format: 'plain-text',
  length: 'medium'
});

const result = await summarizer.summarize(selectedText);
Configuration options:
  • Type: tldr, key-points, teaser, headline
  • Format: plain-text, markdown
  • Length: short, medium, long

Translate

Language detection
Speak any language with Chrome’s Translator API and Language Detector API:
  • Translate to 100+ languages instantly
  • Auto-detect source language
  • Works completely offline
  • No external API calls required
// Auto-detect source language
const detector = await self.translation.createDetector();
const results = await detector.detect(text);
const sourceLang = results[0].detectedLanguage;

// Translate to target language
const translator = await self.translation.createTranslator({
  sourceLanguage: sourceLang,
  targetLanguage: 'es'
});

const translated = await translator.translate(text);
Supported languages include:
  • English, Spanish, French, German, Italian, Portuguese
  • Chinese, Japanese, Korean
  • Arabic, Hindi, Russian
  • And 100+ more languages

Rewrite

Make your words shine with Chrome’s Rewriter API:

Fix Grammar

Clean up spelling and grammar mistakes

Change Tone

Go formal, casual, or professional

Adjust Length

Make it shorter, expand it, or keep it concise

Improve Clarity

Simplify complex sentences
const rewriter = await self.aiRewriter.create({
  tone: 'formal',
  format: 'as-is',
  length: 'as-is'
});

const rewritten = await rewriter.rewrite(text, {
  context: 'Professional email to a client'
});
Tone options:
  • as-is - Keep original tone
  • formal - Professional and polished
  • casual - Friendly and conversational
Length options:
  • as-is - Keep original length
  • shorter - Make it more concise
  • longer - Expand with more detail

Dictionary

Your personal word expert using Chrome’s Prompt API (Gemini Nano):

Definitions

Get clear, concise definitions

Synonyms

Find alternative words

Pronunciation

Learn how to say it

Examples

See real usage in context

Writer

Generate fresh content from scratch using Chrome’s Writer API:
  • Create content based on your ideas
  • Works with selected text as context
  • Perfect for brainstorming and drafting
  • Specify tone, format, and length
const writer = await self.aiWriter.create({
  tone: 'neutral',
  format: 'plain-text',
  length: 'medium'
});

const content = await writer.write(
  'Write a brief introduction about AI assistants',
  { context: selectedText }
);

Image Tools

Hover over any image on a webpage to access AI-powered image analysis:

Describe

Multimodal Prompt APIAI tells you what’s in the image with detailed descriptions

Extract Text

OCRPull out text from screenshots, photos, and documents

Identify Objects

Object DetectionSpot and label things in photos automatically
// Describe image content
const session = await self.LanguageModel.create({
  temperature: 1.0,
  topK: 3
});

const description = await session.prompt(
  'Describe this image in detail',
  { image: imageElement }
);

Voice Dictation

Voice dictation
Talk instead of type. The Multimodal Input API lets you dictate directly into text fields:
  • Click the microphone icon in any text input
  • Speak naturally
  • Watch your words appear in real-time
  • Works completely on-device for privacy
const session = await self.LanguageModel.create({
  temperature: 0.1,
  topK: 3,
  outputLanguage: 'en'
});

const transcription = await session.prompt(
  'Transcribe this audio',
  { audio: audioBlob }
);
Voice dictation requires the Multimodal Input Chrome flag to be enabled. See Installation for setup instructions.

How It Works

The AI Toolbar uses Chrome’s Built-in AI APIs for all processing:
1

Text Selection

When you select text or focus an input field, VAssist detects the context and displays relevant tools
2

On-Device Processing

All AI operations run locally using Gemini Nano - no data leaves your device
3

Instant Results

Results appear in a floating panel near your selection with options to copy, edit, or apply

Document Interaction

Beyond text and images, the toolbar provides smart document features:
  • Page Context - Ask questions about the current page
  • Smart Summaries - Get instant summaries of articles and documents
  • Content Analysis - Understand complex content with AI assistance
All toolbar features work seamlessly across any website without requiring additional permissions or external API keys.

Configuration

Customize toolbar behavior in the Control Panel:
  • Enable/disable specific tools
  • Set default languages for translation
  • Configure summarizer length and format
  • Choose rewriter tone preferences
See Configuration for detailed settings.

Privacy & Performance

100% Local

All processing happens on your device. No data sent to external servers.

Fast & Efficient

Chrome’s optimized Gemini Nano model provides instant results.

No API Keys

Works out of the box without any API keys or accounts.

Offline Ready

Most features work completely offline after initial model download.

Next Steps

Chat Interface

Learn about the conversational chat features

Chrome AI APIs

Deep dive into Chrome’s built-in AI capabilities

Build docs developers (and LLMs) love