Axon - Screen Assistant

Instantly analyze your screen, summarize text, and ask questions about any webpage using advanced Vision AI.

# Axon: The Cognitive Browser Assistant **Axon** is an intelligent Chrome, Firefox & Edge extension that acts as your browser's "Digital Twin". It integrates state-of-the-art AI (Llama-3 via Groq) to see what you see, understand your context, and automate tedious web tasks. ## 🚀 Features ### 👁️ **Vision Analysis** Axon can "see" your current tab. - **"Look at this screen"**: Instantly captures the visible area and lets you ask questions about charts, images, or designs. - **Contextual Help**: Ask "What does this error mean?" or "Summarize this dashboard". ### 📝 **Smart Autofill** Forget rigid form fillers. Axon understands natural language. - **Command**: *"Put my work email in the subscription box"* or *"Fill this application with my dev persona"*. - **Logic**: It scans the page, maps your stored profile data to the correct fields, and fills them intelligently. ### 🧠 **Adaptive Memory** Axon learns as you chat. - If you Mention: *"My phone number is 555-0199"*, Axon saves it to your encrypted profile. - Next time you ask to fill a form, it remembers. ### 💬 **In-Page Chat (Glass & Gradient)** A beautiful, non-intrusive chat overlay that floats above your web page. - **Premium UI**: Glassmorphism aesthetic with smooth animations. - **Control**: Minimize, maximize, or close at any time. ## 🛠️ Installation (Developer Mode) 1. **Clone this repository**: ```bash git clone https://github.com/yourusername/axon-assistant.git ``` 2. **Open Chrome Extensions**: - Navigate to `chrome://extensions/` - Toggle **Developer mode** (top right). 3. **Load Extension**: - Click **Load unpacked**. - Select the folder where you cloned this repo. ## ⚙️ Configuration 1. Click the **Axon Icon** in your toolbar. 2. Select **Settings** (Gear Icon). 3. **API Key**: Enter your Groq API Key (supports Llama-3-70b). - *Don't have one? Get it free at [console.groq.com](https://console.groq.com)* 4. **Profile**: Add your Role, About Me, and any custom Key-Value facts (e.g., Portfolio URL, LinkedIn) to power the Autofill engine. ## ⌨️ Usage - **Right-Click Menu**: Quick access to "Axon: Assist" or "Axon: Autofill". - **Popup**: Click the icon to launch the Control Panel. - **Commands**: - *"Analyze this page"* -> Triggers Vision Agent. - *"Fill formatting form"* -> Triggers Form Agent. - *"Explain this code"* -> Triggers Chat Agent. ## 🏗️ Architecture - **Frontend**: Vanilla JS, HTML5, CSS3 (No framework bloat). - **Design System**: Custom CSS variables for consistent Glass/Gradient theming (`design.css`). - **Backend**: `background.js` acts as an Agent Orchestrator. - **Router**: Classifies intent (Vision vs. Form vs. Chat) using a specialized LLM prompt. - **Agents**: Modular classes (`ChatAgent`, `VisionAgent`, `FormAgent`) handling specific domains. --- *Built with ❤️ for the future of browsing.*