🔮 Glimpse

Intelligent AI assistant sidebar with YouTube transcript analysis, screenshot capture, auto-attach active tab content, interactive help, and file upload support using Gemini and OpenRouter APIs. Now supports Google Search grounding (citations) for Gemini.

✨ Features

• 🤖 AI Chat Interface - Clean, modern sidebar interface for AI conversations
• 📺 YouTube Transcript Analysis - Fetch and analyze YouTube video transcripts
• 📸 Screenshot Capture - Capture and analyze screenshots with AI
• 📄 Tab Content Capture - Extract and analyze content from web pages
• 🔗 Auto‑include Active Tab - Optionally auto‑attach the current tab’s main content when sending
• 📁 File Upload Support - Upload and analyze documents, images, and more
• 🔑 Multiple AI Providers - Support for Gemini and OpenRouter APIs
• 🌐 Google Search Grounding - Let Gemini consult the web and add inline citations (toggle in Settings)
• 💾 Chat History - Save and manage conversation history locally
• ⚙️ Configurable Settings - Customize API keys, models, and preferences
• 🆘 Built‑in Tutorial - Help button opens a step‑by‑step guide with interactive actions
• 💬 @Mentions for Tabs - Type @ to attach open tabs as context
• 🔁 Copy/Retry - Copy any message and retry AI responses easily
• 🧱 Rich Code Blocks - Clean formatting with copy buttons
• 🧑‍💼 Personalization (Profile) - Add “About me,” attach .txt/.md/.pdf files, auto‑generate a compact profile summary, and choose whether to include it by default or only for the next message
• 🔐 Privacy First - All data stored locally, no analytics or tracking

🚀 Installation

🦊 Firefox Add-ons Store (Recommended)

📥 Install from Mozilla Add-ons

The easiest way to install Glimpse is directly from the official Firefox Add-ons store.

Manual Installation

Method 1: Temporary Installation (Development)

1. Open Firefox and navigate to about:debugging
2. Click "This Firefox" in the left sidebar
3. Click "Load Temporary Add-on..."
4. Select the manifest.json file from this repository
5. The extension will be loaded until Firefox restarts

Method 2: Manual XPI Installation

1. Download or clone this repository
2. Run package-simple.ps1 to create the .xpi file
3. Drag and drop the .xpi file into Firefox
4. Confirm installation when prompted

Chrome/Edge (Experimental)

While primarily designed for Firefox, basic functionality may work in Chrome:

1. Open chrome://extensions/
2. Enable "Developer mode"
3. Click "Load unpacked" and select the repository folder

⚙️ Setup

1. Install the extension using one of the methods above
2. Open the sidebar by pressing F1 or Ctrl+Shift+Y
3. Configure API keys in the settings:
   • Gemini API: Get your free API key from Google AI Studio
   • OpenRouter API: Get your API key from OpenRouter
4. Optional: Enable Google Search grounding to get real‑time citations
5. Optional: Enable Auto‑include active tab to attach the current page content automatically
6. Start chatting with your AI assistant!

Keyboard & Quick actions

• F1 or Ctrl+Shift+Y: toggle sidebar
• Click the logo: start a new session
• Enter to send, Shift+Enter for newline
• Use the toolbar to Capture Page, Screenshot, YouTube Analysis, or Attach Files

🛠️ Development

Quick Start

1. Clone the repository

   git clone https://github.com/Zandaland/glimpse.git
   cd glimpse

2. Make changes to any source files

3. Package the extension

   .\package-simple.ps1

4. Test in Firefox by dragging the new .xpi file into the browser

File Structure

glimpse/
├── manifest.json      # Extension configuration
├── background.js      # Background processes and API
├── sidebar.html       # Main UI structure
├── sidebar.css        # Main styling
├── sidebar.js         # Core functionality and AI logic
├── retry.css          # Additional styling
├── transcript-server/ # Optional helper for YouTube transcripts (Docker/Python)
├── icons/             # Extension icons
├── package-simple.ps1 # Build script
└── README.md          # This file

Key Functions (sidebar.js)

• sendMessage() - Handles user input and AI responses
• callGeminiAPI() - Makes API calls to AI services
• captureTabAsFile() / captureCurrentTab() - Extract main content from active tab
• analyzeYouTubeVideo() - YouTube transcript extraction
• captureScreenshot() - Screenshot functionality
• addMessage() - Message rendering and display

Google Search Grounding

• Toggle “Enable Google Search grounding” in Settings. For Gemini 2.x it uses google_search; for Gemini 1.5 it falls back to legacy dynamic retrieval. Responses include inline citations when provided.

Auto‑include Active Tab

• Toggle “Auto‑include active tab when sending.” The extension captures the main content from the current tab and attaches it as a “Webpage” file (skips internal pages like about:). This works like the Capture Page button, but automatically.

Personalization (Profile)

Glimpse can personalize responses using a lightweight profile you control.

• Open Settings → “Personalization (Profile)”
  • Write your “About me” details (interests, goals, writing style, do/don’ts)
  • Optionally attach reference files (.txt, .md, .pdf)
  • Click “Generate summary” to create a compact profile summary that the model can use efficiently
  • Choose how to include your profile:
    • “Include by default” — always include in prompts
    • “Include for the next message only” — one‑off include
• Your profile is stored locally in the browser. You can edit or remove it anytime.
• When included, Glimpse prepends a small [User Profile] context block to your conversation turn.

🔒 Privacy

Glimpse is designed with privacy in mind:

• Local Storage Only - All settings, API keys, and chat history are stored locally in your browser
• No Analytics - We don't collect any usage data or analytics
• Direct API Calls - Your data goes directly to your chosen AI provider (Gemini/OpenRouter)
• No Cloud Storage - We don't store any of your data on our servers

For full details, see our Privacy Policy.

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request. For major changes, please open an issue first to discuss what you would like to change.

📝 License

This project is licensed under the MIT License - see the LICENSE file for details.

🐛 Issues & Support

If you encounter any issues or have suggestions, please open an issue on GitHub.

---

Made with ❤️ by Zandaland