Self-Hosted AI Stack

A complete guide to running your own AI stack on local hardware — chat models, image generation, IDE coding agent, live web search, and secure remote access. No API keys, no subscriptions.

Part 1 of the self-hosted AI stack series. Understand what actually determines what you can run — VRAM budgets, model sizing, and the hardware decisions that affect everything else.

Part 2 of the self-hosted AI stack series. Get a dedicated WSL2 distro running on the drive of your choice, with Docker Engine and GPU passthrough confirmed working before installing a single AI service.

Part 3 of the self-hosted AI stack series. Get Ollama, Open WebUI, and SearXNG running together as a single Docker Compose stack — the foundation everything else builds on.

Part 4 of the self-hosted AI stack series. Get ComfyUI generating images end-to-end, understand the node graph well enough to debug it yourself, and learn the failure modes that come up with multi-subject compositions.

Part 5 of the self-hosted AI stack series. Wire your locally-hosted coding model into VS Code as a real agentic assistant — not just chat, but something that reads your open file, edits code in place, and reasons about actual bugs.

Part 6 of the self-hosted AI stack series. Reach your stack securely from outside your home network — including from behind CGNAT, where traditional port forwarding isn't an option.

Part 7 of the self-hosted AI stack series. A debugging method for when something's wrong and you're not yet sure which layer it's in — plus a quick index pointing back to specific fixes in earlier parts.