Show HN: [GPU-Pro] Master Your AI Workflow

GPU Pro is the modern solution for NVIDIA GPU monitoring, designed for AI engineers, ML researchers, and GPU cluster administrators. Get real-time insights into your GPU infrastructure with:

• 🎨 Beautiful Web UI - Sleek, responsive dashboard with real-time updates
• 💻 Terminal UI - Elegant TUI for SSH sessions and headless servers
• 📊 Comprehensive Metrics - GPU utilization, memory, temperature, power, processes, and more
• 🌐 Network Monitoring - Track connections, bandwidth, and geolocation
• 💾 System Insights - CPU, RAM, disk, and fan monitoring
• 🔌 Hub Mode - Aggregate multiple GPU nodes into one dashboard
• 🚀 Zero Setup - Single binary, no Python, no Node.js, no containers required
• ⚡ Lightning Fast - Built with Go for maximum performance

• Real-time Metrics: Utilization, memory usage, temperature, power consumption
• Process Tracking: See what's running on each GPU with detailed process information
• Multi-GPU Support: Monitor all GPUs simultaneously
• Historical Charts: Track performance over time with customizable time ranges
• NVML & nvidia-smi: Supports both native NVML library and nvidia-smi fallback

• Network I/O: Real-time bandwidth tracking with historical charts
• Disk I/O: Monitor read/write operations and throughput
• Network Connections: Live connection tracking with geolocation on world map
• Open Files: Track open file descriptors and large files
• System Resources: CPU, RAM, disk usage, and fan speeds

• Web Dashboard: Modern, responsive design with glassmorphism effects
• Terminal UI (TUI): Beautiful colored terminal interface for SSH sessions
• Dark Theme: Professional tech-inspired design optimized for long sessions
• Real-time Updates: WebSocket-based live data streaming
• Mobile Responsive: Works perfectly on phones and tablets

• Single Binary: Everything embedded - templates, assets, and code
• Cross-Platform: Linux, Windows, macOS support
• Multiple Modes: Standalone or hub aggregation mode
• Easy Deployment: Systemd service, Docker, or bare metal
• Zero Dependencies: Just NVIDIA drivers required

Stunning real-time GPU monitoring with modern design

https://youtu.be/vwBC54nXOPI

Elegant TUI for SSH and headless servers

The fastest way to get started:

or

That's it! The script will run the project and let you choose between Web UI or Terminal UI.

• NVIDIA GPU with drivers installed
• Go 1.24+ (for building from source)
• Linux, Windows, or macOS

Access the dashboard at http://localhost:1312 (or your custom port).

Perfect for SSH sessions and headless servers!

All configuration is done via environment variables:

• AI/ML Training: Monitor GPU utilization during model training
• Research Labs: Track multi-GPU workstations and servers
• GPU Clusters: Aggregate monitoring across multiple nodes
• Cloud GPU Instances: Monitor AWS, GCP, or Azure GPU VMs
• Gaming Rigs: Track GPU performance during gaming sessions
• Crypto Mining: Monitor mining rig performance and temperatures
• Remote Work: SSH-friendly TUI for remote GPU monitoring

Contributions are welcome! Here's how you can help:

1. 🐛 Report Bugs: Open an issue with detailed reproduction steps
2. 💡 Suggest Features: Share your ideas in the issues section
3. 🔧 Submit PRs: Fork, create a feature branch, and submit a pull request
4. 📖 Improve Docs: Help us make the documentation better
5. ⭐ Star the Project: Show your support!

MIT License - see LICENSE file for details

• NVIDIA for the NVML library
• Go NVML bindings by NVIDIA
• Charm for beautiful TUI components (Bubble Tea, Lipgloss)
• Fiber for blazing-fast web framework
• Chart.js for stunning data visualizations