J.A.R.V.I.S.

About

# J.A.R.V.I.S. AI Assistant
## Comprehensive Feature Documentation

J.A.R.V.I.S. is an advanced AI-powered Chrome extension that provides a floating assistant with smart voice commands and contextual responses.

## Popup Menu Features

The extension settings popup offers the following configuration options:

### Core Settings
- **Enable/Disable Assistant** - Toggle switch to turn the AI assistant on or off across all websites
- **AI Model Selection** - Choose from multiple AI model options:
  - llama3-70b-8192
  - llama-3.3-70b-versatile
  - gemma2-9b-it
  - deepseek-r1-distill-llama-70b
- **Theme Selection** - Customize appearance with options:
  - Light
  - Dark
  - Auto (Follow System)
- **Widget Position** - Control where the floating widget appears:
  - Bottom Right
  - Bottom Left
  - Top Right
  - Top Left
- **Save Settings Button** - Applies and stores your configuration preferences

## Floating Chat Box Capabilities

### Basic Interface
- **Draggable Widget** - The 🤖 icon can be positioned anywhere on screen
- **Text Input** - Type questions or commands in the chat box
- **Message History** - Maintains conversation history in the chat interface
- **Smart Contextual Responses** - AI considers the content of the current webpage

### Control Buttons
- **Wake Mode Button** (👂) - Activates "wake word" listening mode
- **Voice Chat Button** (🎤) - Enables voice recognition for hands-free interaction
- **Clear Chat Button** (🗑️) - Erases the conversation history
- **Close Button** (×) - Minimizes the chat back to the widget icon

### Voice Interaction
- **Wake Word Detection** - Say "Jarvis" to activate the assistant
- **Continuous Voice Mode** - Conduct conversations entirely by voice
- **Voice Indicator** - Visual feedback showing when the assistant is listening
- **Voice Synthesis** - Natural-sounding speech responses

### AI Capabilities
- **Context-Aware Responses** - Considers page content for relevant answers
- **Command Detection** - Intelligently processes questions vs. commands
- **Markdown Formatting** - Supports code blocks, lists, bold/italic text
- **Task Specialization** - Optimized for various tasks:
  - Summarization
  - Code help
  - Explanations
  - Reviews
  - Translation
  - Rewriting

### Browser Control Commands

#### Navigation
- **"Go to [website]"** - Navigate to a specific website
- **"Search for [query]"** - Perform a Google search
- **"Back" or "Go back"** - Navigate to the previous page
- **"Forward" or "Go forward"** - Navigate to the next page

#### Page Interaction
- **"Click [element]"** - Click on buttons, links, or other elements
- **"Fill [form field] with [value]"** - Enter text in forms
- **"Scroll down/up"** - Scroll the page in either direction
- **"Scroll to top/bottom"** - Jump to beginning or end of page
- **"Scroll to [section]"** - Find and scroll to specific content

#### Browser Controls
- **"Refresh" or "Reload"** - Refresh the current page
- **"Bookmark" or "Save"** - Bookmark the current page
- **"Close tab" or "New tab"** - Tab management
- **"Screenshot" or "Capture"** - Take a screenshot of the current page

#### Media Controls
- **"Play/Pause/Stop"** - Control video or audio playback
- **"Mute/Unmute"** - Toggle audio
- **"Volume up/down"** - Adjust volume
- **"Skip" or "Next track"** - Move to next media item
- **"Previous track"** - Go to previous media item

#### Utility Actions
- **"Clear chat"** - Clear the conversation history
- **"Extract [content]"** - Pull specific information from the page

### Technical Features
- **Tab-Based Activation** - Voice recognition pauses when tab is inactive
- **Cross-Site Functionality** - Works across all websites with permissions
- **Persistence** - Maintains settings across browser sessions
- **Error Handling** - Graceful recovery from API or processing failures
- **Context Management** - Maintains conversation history for contextual responses

## Content Analysis
- **Page Content Extraction** - Identifies and extracts main content
- **Metadata Reading** - Accesses page metadata for better context
- **Code Block Detection** - Special handling for pages with code content
- **Smart Element Finding** - Locates elements based on descriptions

---

J.A.R.V.I.S. effectively combines a powerful AI chat interface with browser automation capabilities, allowing users to interact with both content and browser controls through natural language commands.