🤖 A private LLM in Docker, in minutes

Ollama + LiteLLM + Qdrant + Open WebUI, wired together and production-hardened. Clone, make up, and you have private chat + RAG + an OpenAI-compatible API. Your data never leaves your machine.

Get the kit — from $14.50 Get the free Lite version on GitHub

One-time purchase · runs on your own server · CPU or GPU

                       ┌──────────────┐
   you ──HTTPS──▶ Traefik ─▶ Open WebUI │  chat · accounts · RAG
                       └──────┬───────┘
                  ┌───────────┼───────────┐
                  ▼           ▼           ▼
              Ollama      LiteLLM      Qdrant
           local models  OpenAI-API   vector store
                            └─▶ (optional) OpenAI / Anthropic

Private by default

Models run locally with Ollama; the UI binds to localhost until you turn on HTTPS. Nothing phones home.

Actually wired together

Not four containers in a file — Open WebUI talks to Ollama and LiteLLM, uses Qdrant for RAG, with cloud fallback if you want it.

Production-hardened

One-command Let’s Encrypt HTTPS, GPU overlay, backups, healthchecks, and secure signup defaults — the parts homelab guides skip.

Pricing

One-time purchase, yours to deploy anywhere

Start free on your laptop, upgrade when you put it on a server.

Lite

Free

The full stack on localhost. Great for trying it on your laptop.

Ollama + LiteLLM + Qdrant + Open WebUI
Local + optional cloud models
Bootstrap + Makefile

Free on GitHub →

Starter

$29$14.50

The full private-AI stack, plus one-command public HTTPS.

Ollama + LiteLLM + Qdrant + Open WebUI, wired together
One-command public HTTPS (Traefik + Let’s Encrypt)
Local models + optional cloud routing with fallback (LiteLLM)
Bootstrap, model-pull, and a full Makefile UX
Commercial-use license (single organization)

Buy Starter — $14.50

Standard

Pro

$79$39.50

Fully operable: monitoring, incident runbooks, updates & support.

Everything in Standard
Monitoring exporters (cAdvisor + node-exporter)
6 incident runbooks tailored to this stack
One year of updates
Priority email support

Buy Pro — $39.50

Secure Stripe checkout · instant download · commercial-use license

Included

What you get

Ollama local models
OpenAI-compatible API (LiteLLM)
RAG vector search (Qdrant)
Open WebUI chat
Traefik + Let’s Encrypt HTTPS
NVIDIA GPU acceleration
Backup & restore
Monitoring & incident runbooks

Questions

FAQ

Does my data leave my server?

No. Ollama runs the models locally and everything is stored on the host you deploy to. Cloud model fallbacks (OpenAI/Anthropic) are optional and off unless you add a key.

Do I need a GPU?

No — it runs on CPU out of the box (smaller models are fine). The Standard and Pro tiers add an NVIDIA GPU overlay for acceleration when you have one.

What exactly am I buying?

The integration: tested Docker Compose files, overlays, a Makefile, bootstrap/backup scripts, and documentation. The underlying tools (Ollama, Open WebUI, Qdrant, LiteLLM, Traefik) are open source and keep their own licenses.

How do I get updates?

Re-download from your purchase link. Pro includes one year of updates and priority email support.

Can I use it for client work?

Yes. The commercial license covers deploying it for yourself, your employer, or your clients. You may not resell the kit itself.

Try it free, then make it production

Download Lite and run it on your laptop in three commands. When you're ready to put it on a server with HTTPS, GPU, and backups, upgrade.

Get free Lite on GitHub See the tiers