Under the hood

Every tool that keeps the Workshop running, each with its job in a line. Open source where it counts, hardened at the edge, and watched by its own monitors. None of it is exotic, and that is the point: these are off-the-shelf parts, assembled with care.

The AI

Local models run on the machine in the room, so private work stays private. The leading frontier models sit one switch away through the same gateway, for when a task needs more horsepower.

Open WebUI

Chat front end for every model, local or frontier

Ollama

Runs the language models locally on the platform

LiteLLM

One private gateway in front of every model

Claude

Anthropic’s frontier models, reached through the gateway

ChatGPT

OpenAI’s models, on tap for heavier reasoning

Gemini

Google’s models, available through the same switch

Edge and access

How the Workshop reaches the world without leaving a door open.

Cloudflare

Edge security, Zero Trust Access, and the tunnel

Caddy

Reverse proxy and automatic HTTPS

CrowdSec

Detects attacks and bans IPs at the edge

Tailscale

Private mesh VPN back to the platform

Security

This is a security professional’s home lab, so the defenses are the work, not an afterthought. Protection is layered at the edge and on the host. The live internals, the counts of blocked attempts and banned addresses, stay private on purpose. What is shown here is the posture, not the playbook.