Run Your Own AI Image Generator at Home with Docker and Open WebUI

Imagine creating stunning AI-generated images right on your own computer—no cloud subscriptions, no privacy concerns, no credit counters. With Docker Model Runner and Open WebUI, you can set up a fully local image generation pipeline in just a few minutes. This guide walks you through everything you need, from what the tool does to how to generate your first dragon in a business suit. Below, we answer your burning questions about this powerful, private setup.

What exactly is Docker Model Runner and how does it work with Open WebUI?

Docker Model Runner is a lightweight control plane that manages the entire lifecycle of AI models on your local machine. It downloads image generation models, spins up the necessary inference backends, and exposes an API that is 100% compatible with OpenAI's endpoints—including the POST /v1/images/generations pathway. Open WebUI, a popular chat interface for AI tools, is preconfigured to talk to this exact API. When you launch Open WebUI via Docker Model Runner, the two connect automatically: you send a prompt from the chat window, Docker Model Runner processes it through the locally stored model, and the generated image appears right in the UI. All data stays on your computer, never leaving your network.

Run Your Own AI Image Generator at Home with Docker and Open WebUI — Source: www.docker.com

What hardware and software do I need to get started?

The requirements are modest. You need either Docker Desktop (on macOS) or Docker Engine (on Linux). For a small image generation model, allocate at least 8 GB of free RAM—more RAM improves performance and allows larger models. A GPU is optional but highly recommended: NVIDIA GPUs with CUDA, Apple Silicon with MPS, or even a CPU fallback all work. The Docker Model Runner will automatically use the best available hardware. To verify your setup, run docker model version in your terminal. If you see version info without errors, you are ready to proceed.

How do I download an image generation model locally?

Docker Model Runner uses a compact packaging format called DDUF (Diffusers Unified Format) to distribute models through Docker Hub, just like any other OCI artifact. To fetch a model, open your terminal and run:
docker model pull stable-diffusion
This command downloads the latest Stable Diffusion XL base model (about 6.94 GB) to your local machine. You can confirm the download with:
docker model inspect stable-diffusion
The output shows the model's SHA256 hash, tags, and details like the DDUF file name and layout. Under the hood, the DDUF file bundles all components—text encoder, VAE, UNet/DiT, and scheduler configuration—into a single portable artifact that Docker Model Runner unpacks during inference.

How do I launch Open WebUI and connect it to the local model?

This is the simplest step. Docker Model Runner includes a built-in launch command that automatically wires everything together. Just run:
docker model launch openwebui
That single command starts the Open WebUI container and configures it to use your locally running model's API endpoint. After a few seconds, you'll see a URL (typically http://localhost:3000) in your terminal. Open that in your browser, and you'll be greeted by a chat interface ready to generate images. No manual API key setup, no complex networking—Docker Model Runner handles all the wiring behind the scenes.

Source: www.docker.com

How do I actually generate an image from the chat interface?

Once Open WebUI is loaded, simply type your prompt into the chat input. For example, you could type: “A dragon wearing a business suit, sitting in a boardroom, photorealistic” and press Enter. The message is sent to Docker Model Runner, which runs it through the Stable Diffusion model on your machine. Within seconds, the generated image appears in the conversation. You can continue refining your prompt, generate variations, or ask for tweaks—all without any cloud dependency. Since the model runs locally, there are no content filters beyond what you choose, no rate limits, and no per-image costs.

What are the main advantages of running image generation locally?

Running everything on your own machine gives you complete privacy—your prompts and generated images never leave your computer. You also avoid subscription fees, credit systems, and usage caps that come with cloud services. Because you control the model and settings, you can use any prompt you like without worrying about arbitrary content restrictions. Additionally, local inference often exhibits lower latency once the model is loaded, and you can experiment with different models or tweak parameters freely. The setup is also portable: once you have Docker Desktop or Engine, you can replicate the environment on any compatible machine.

Can I customize the experience or use different models?

Absolutely. Docker Model Runner supports a range of DDUF-packaged image generation models available on Docker Hub. You can pull alternative Stable Diffusion versions or other diffusion models by changing the model name in the docker model pull command. Within Open WebUI, you can adjust generation parameters like image size, number of steps, and guidance scale if you extend the interface with plugins or modify the configuration. For advanced users, you can even run multiple models side by side by launching separate instances or using Docker Compose. The local environment is your sandbox—feel free to experiment without risking subscription costs or data leaks.

Tags: