ollvision/README.md

Ollama Vision Chat 🎥💬

A real-time camera-enabled chat application that connects to your local Ollama vision model, allowing you to have conversations with an AI that can see and understand what you're doing through your camera.

✨ Features

🎥 Real-time Camera Integration

• Continuous camera feed with live preview
• Automatic frame capture at configurable intervals (5-60 seconds)
• Manual photo capture with instant feedback
• Pause/resume auto-capture functionality
• Mobile and desktop camera support

💬 Interactive AI Chat

• Real-time messaging with vision-enabled AI
• Context-aware responses based on camera feed
• Message history with clear user/assistant distinction
• Auto-resizing text input for comfortable typing
• Smooth animations and modern UI

⚙️ Flexible Configuration

• Configurable Ollama server URL
• Support for multiple vision models (LLaVA, etc.)
• Adjustable capture intervals
• Real-time connection status monitoring
• Easy settings panel with instant updates

📱 Responsive Design

• Optimized for desktop and mobile devices
• Adaptive layout that works on any screen size
• Touch-friendly controls
• Modern glassmorphism design with smooth animations

🚀 Quick Start

Prerequisites

1. Install Ollama

   # On macOS
   brew install ollama
   
   # On Linux
   curl -fsSL https://ollama.ai/install.sh | sh
   
   # On Windows
   # Download from https://ollama.ai/download
   

2. Pull a Vision Model

   # LLaVA (recommended)
   ollama pull llava
   
   # Or other vision models
   ollama pull llava:7b
   ollama pull llava:13b
   ollama pull bakllava
   

3. Start Ollama Server

   ollama serve
   

Installation & Usage

1. Download the HTML file or clone this repository
2. Open index.html in a modern web browser
3. Grant camera permissions when prompted
4. Wait for connection - the status indicator will turn green
5. Start chatting! The AI can see what you're doing and respond accordingly

🛠️ Configuration

Settings Panel

Click the ⚙️ settings button to access:

• Ollama Server URL: Default http://localhost:11434
• Vision Model: Default llava (change to your preferred model)
• Auto-capture Interval: 5-60 seconds between automatic photos
• Enable Auto-capture: Toggle automatic photo capturing

Supported Models

The app works with any Ollama vision model:

• llava - General purpose vision model
• llava:7b - Smaller, faster version
• llava:13b - Larger, more capable version
• bakllava - Alternative vision model
• Custom vision models you've imported

💡 Usage Examples

What You Can Ask:

• "What am I holding in my hands?"
• "How does this setup look?"
• "Can you see what's on my screen?"
• "What color is my shirt?"
• "Help me organize this workspace"
• "What do you think of this drawing?"
• "Can you read this text for me?"

Perfect For:

• 🎨 Getting feedback on artwork or projects
• 📚 Reading assistance and text recognition
• 🏠 Home organization and decoration advice
• 👔 Fashion and styling feedback
• 🔧 Technical troubleshooting with visual context
• 🍳 Cooking assistance and recipe guidance
• 📖 Educational support with visual learning

🔧 Technical Details

Architecture

• Frontend: Pure HTML5, CSS3, and JavaScript (no frameworks)
• Camera: WebRTC getUserMedia API for camera access
• Image Processing: HTML5 Canvas for frame capture
• Communication: Fetch API for Ollama REST API calls
• UI: Modern CSS Grid and Flexbox layouts

Browser Requirements

• Modern browser with camera support
• JavaScript enabled
• HTTPS or localhost (required for camera access)
• WebRTC support

File Structure

ollama-vision-chat/
├── index.html          # Main application file
├── README.md          # This file
└── screenshots/       # (Optional) App screenshots

🚨 Troubleshooting

Common Issues & Solutions

🔴 "Failed to connect to Ollama"

• Ensure Ollama is running: ollama serve
• Check the server URL in settings (default: http://localhost:11434)
• Verify your firewall isn't blocking the connection

📷 "Failed to access camera"

• Grant camera permissions in your browser
• Ensure no other app is using the camera
• Try refreshing the page and granting permissions again
• Use HTTPS or localhost (camera requires secure context)

🤖 "Model not found"

• Install a vision model: ollama pull llava
• Check available models: ollama list
• Update the model name in settings to match your installed model

⏱️ "Slow responses"

• Try a smaller model like llava:7b
• Increase the auto-capture interval to reduce processing load
• Ensure your system has adequate RAM and CPU

📱 "Mobile camera issues"

• Grant camera permissions in browser
• Try both front and back cameras
• Ensure mobile browser supports WebRTC

🎯 Performance Tips

• Use smaller models (llava:7b) for faster responses
• Adjust capture intervals based on your use case
• Pause auto-capture when not needed to save resources
• Close other camera applications to avoid conflicts
• Use a stable internet connection for best performance

🔒 Privacy & Security

• All processing is local - your camera feed never leaves your device
• No data is stored - conversations are not saved
• Direct Ollama connection - no third-party services involved
• Open source - you can review all code in the HTML file

🤝 Contributing

Contributions are welcome! Here are some ideas:

• 🎨 UI/UX improvements
• 🔧 Additional model support
• 📱 Enhanced mobile experience
• 🌐 Multi-language support
• 📊 Usage analytics dashboard
• 🔄 Conversation export/import

📋 Roadmap

• Voice input/output support
• Multiple camera sources
• Image annotation tools
• Conversation export
• Custom model training integration
• Screen sharing capability
• Multi-user support

📜 License

This project is open source and available under the MIT License.

⭐ Support

If you find this project helpful:

• ⭐ Star this repository
• 🐛 Report issues and bugs
• 💡 Suggest new features
• 🤝 Contribute improvements

---

Built with ❤️ for the Ollama community

Have questions? Open an issue or start a discussion!