Complete Setup Guide
Build Your Own Local AI Workstation
A full AI stack running on your own hardware — chat, code, image generation, video generation, and speech. Private, unlimited, customizable. No subscription, no cloud, no one watching.
Text & Code
Image Generation
Video Generation
Speech
Security Research
100%
Private
$0
Per query
Why Run AI Locally?
Understanding AI Models
Four distinct types — each does something different.
Best models:
Hardware Requirements
The GPU is the most important component — it determines which models you can run and at what speed.
- —
⚠️
VRAM is the bottleneck. A model must fit entirely in GPU memory to run at full speed. Running on CPU only is possible but 10–50× slower. For video generation (Wan 2.x, Hunyuan), plan for 16–24 GB VRAM minimum.
The Recommended Stack
A complete local AI suite covering every modality.
| Purpose | Software | Model | VRAM | Notes |
|---|---|---|---|---|
Installation Guide
Step by step. Follow in order — later steps depend on earlier ones.
💡
What You Can Do With It
Practical use cases once everything is running.
- →
Supercharge with MCP Servers
Model Context Protocol lets your local AI read files, call APIs, query databases, and use tools — turning a chat bot into an actual agent.
| MCP Server | Gives the AI access to | Install |
|---|---|---|
Add to your Open WebUI or Ollama MCP config:
{
"mcpServers": {
"filesystem": {
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-filesystem", "/path/to/your/repos"]
},
"github": {
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-github"],
"env": { "GITHUB_TOKEN": "your_token_here" }
}
}
}