One command. Full AI stack.
On your Ubuntu server.
AIStack is an open-source Go binary that detects your GPU, installs Ollama + Open WebUI + llama-swap, and hands you a running AI inference stack in minutes.
// What it does
Auto-detects GPU
Detects NVIDIA CUDA GPUs (sm_60+) and configures optimal settings for your hardware automatically.
Installs Ollama
Downloads and installs Ollama with GPU acceleration enabled. No manual CUDA toolkit setup needed.
Sets up Open WebUI
Deploys a clean web interface for chatting with local models. Accessible at localhost:3000 out of the box.
Configures llama-swap
Multi-model lazy loading so you can switch between models without restarting the whole stack.
Systemd services
Generates systemd unit files for auto-start on boot. Your AI stack runs reliably as a background service.
Ubuntu 22.04 / 24.04
Tested and supported on the two latest Ubuntu LTS releases. Works on bare metal and in the cloud.
// Why AIStack
Setting up a local AI stack usually takes hours of config and searching Stack Overflow. AIStack does the hard part.
No Docker required. No YAML archaeology. Just run it.
Need a managed setup or custom AI infrastructure?