🦙 Ollama Tray App

A lightweight Linux system tray application to monitor and control the ollama model server.

✅ Features

• Start/stop the ollama systemd service
• Monitor loaded models
• View GPU memory usage (NVIDIA only)
• See token context window usage
• Tray notifications for model events
• Open model directory in file manager
• NEW: Model management (pull/delete models)
• NEW: Run models directly in terminal from the UI
• NEW: Customizable settings via configuration UI
• NEW: About dialog with version information

🧪 Requirements

• Linux (tested on Arch/KDE Wayland)
• Python 3.11+
• Python packages in requirements.txt

🚀 Quick Start

pip install -r requirements.txt
python3 app/ollama-tray.py

📁 Project Structure

ollama-tray/
├── app/
│   ├── config.py        # Configuration management
│   ├── models.py        # Model management utilities
│   ├── ollama-tray.py   # Main application
│   ├── ollama.png       # Application icon
│   ├── ui.py            # UI components for settings and dialogs
│   └── version.py       # Version information
├── appimage/
│   └── AppRun           # AppImage runner script
├── .github/workflows/build-appimage.yml
├── ollama-tray.desktop  # Desktop entry file
├── requirements.txt     # Python dependencies
└── README.md            # Documentation

⚙️ Configuration

The application can be configured via the Settings dialog. Configuration is stored in ~/.config/ollama-tray/config.json.

Available settings:

• Service Name: The systemd service name for ollama (default: ollama.service)
• API URL: The URL of the Ollama API (default: http://127.0.0.1:11434)
• Model Directory: The directory where models are stored (default: ~/.ollama/models)
• Refresh Interval: How often the status is updated (default: 15000 ms)
• Log File: The file where logs are written (default: ~/.cache/ollama-tray.log)
• Log Level: The level of logging detail (default: INFO)

🤖 Model Management

The Model Management feature allows you to:

1. View Installed Models: See all models currently installed in your Ollama instance
2. Pull New Models: Download new models from the Ollama library
3. Remove Models: Delete models you no longer need to free up disk space
4. Monitor Disk Usage: See how much disk space your models are using

Access model management through the tray menu option "Manage Models..."

Available Models

Ollama supports various models including:

• llama2 - Meta's Llama 2 models
• mistral - Mistral AI's models
• gemma - Google's Gemma models
• phi - Microsoft's Phi models
• orca-mini - Smaller, faster models
• And many more!

You can specify model variants like llama2:7b or llama2:13b when pulling.

📝 License

MIT