{"repo":"workhubonline-soft/aistack","free":true,"listed":false,"github":"https://github.com/workhubonline-soft/aistack","clone":"git clone https://github.com/workhubonline-soft/aistack.git","description":"One-command installer that turns a clean Ubuntu server into a fully operational self-hosted AI stack. Automatic hardware detection (CPU/GPU), Docker Compose orchestration, and a built-in Model Advisor that tells you exactly which LLMs will run on your hardware and why.","language":"Go","stars":10,"topics":[],"license":"MIT","category":"self-hosted-apps","readme_excerpt":"AIStack — Self-Hosted AI Server Installer One command to turn a clean Ubuntu server into a fully operational AI stack. What You Get Service URL Description --------- ----- ------------- Open WebUI http://your-server:3000 ChatGPT-like interface Ollama http://localhost:11434 LLM inference engine Nginx (optional) http://your-server:80 Reverse proxy Prometheus + Grafana (optional) :9090 / :3001 Service health monitoring Quick Start One-liner install From source Non-interactive (CI/automation) CLI Reference Model Advisor Example Recommendations by Hardware CPU only (16GB RAM) - ✓ Qwen 3 4B (q4 K M) — best general chat - ✓ Gemma 3 4B (q4 K M) — strong quality for size - ✓ Qwen 3 0.6B (q8 0) — ultra-fast for simple tasks NVIDIA 8GB VRAM - ✓ Qwen 3 8B (q4 K M) — best all-rounder - ✓ Gemma 3 12B (q4 K M) — strong general model - ✓ DeepSeek R1 7B (q5 K M) — best reasoning - ✓ Devstral 24B (q4 K M) — best coding (tight fit) NVIDIA 24GB VRAM - ✓ Qwen 3 32B (q4 K M) — flagship quality - ✓ DeepSeek R1 32B (q4 K M) — top reasoning - ✓ Gemma 3 27B (q5 K M) — Google's best - ✓ Llama 4 Scout (q4 K M) — Meta's MoE (tight fit) Supported Platforms OS Arch CPU NVIDIA GPU ---- ------ ----- ------------ Ubuntu 22.04 LTS x86 64 ✓ ✓ Ubuntu 24.04 LTS x86 64 ✓ ✓ Profiles AIStack auto-detects your hardware and selects a profile: Profile Trigger Ollama config --------- --------- --------------- cpu No GPU detected CPU threads, RAM limit nvidia-low-vram GPU < 8GB VRAM GPU passthrough nvidia-8gb 8–15GB VRAM","default_branch":null,"files":null,"tree":[],"storefront":"/r/workhubonline-soft","claimed":false,"request_supported":{"post":"https://gitbuyer.com/r/workhubonline-soft/aistack/request-supported","requests":0},"note":"indexed from public GitHub; nothing is for sale on this page. Clone it from GitHub. Paid listings live at /search."}