Let\u2019s build it. Your own private DALL-E, no cloud subscription required.<\/em><\/p>\n If you can run docker model version without errors, you\u2019re good to go.<\/p>\n Before we dive in, here\u2019s the big picture:<\/p>\n Docker Model Runner acts as the control plane. It downloads the model, manages the inference backend lifecycle, and exposes a 100% OpenAI-compatible API<\/strong> \u2014 including the POST \/v1\/images\/generations endpoint that Open WebUI already knows how to talk to.<\/p>\n Docker Model Runner uses a compact packaging format called DDUF<\/strong> (Diffusers Unified Format) to distribute image generation models through Docker Hub, just like any other OCI artifact.<\/p>\n Pull a model to get started:<\/p>\n You can confirm it\u2019s ready:<\/p>\n What\u2019s happening under the hood?<\/strong> The model is stored locally as a DDUF file, a single-file format that bundles all the components of a diffusion model (text encoder, VAE, UNet\/DiT, scheduler config) into one portable artifact. Docker Model Runner knows how to unpack it at runtime.<\/p>\n This is a magic trick. Docker Model Runner has a built-in launch command that knows exactly how to wire up Open WebUI against the local inference endpoint:<\/p>\n That\u2019s it. Behind the scenes this runs:<\/p>\n The model-runner.docker.internal hostname is a special DNS entry that Docker Desktop containers use to reach the Model Runner running on the host, no port-forwarding gymnastics required. If you use Docker CE, you\u2019ll see the docker\/model-runner container address instead of model-runner.docker.internal.<\/p>\n Open your browser at http:\/\/localhost:3000<\/strong><\/a>, create a local account (it stays offline), and you\u2019ll land on the chat interface.<\/p>\n Tip:<\/strong> Want to run it in the background? Add \u2013detach:<\/p>\n<\/div>\n Prefer Docker Compose? See the full setup here: https:\/\/docs.docker.com\/ai\/model-runner\/openwebui-integration\/<\/a><\/p>\n Open WebUI already uses Docker Model Runner for text chat automatically (it reads the OPENAI_API_BASE env var). For image generation you need to point it at the images endpoint too, a 30-second job in the settings UI.<\/p>\nWhat You\u2019ll Need<\/h2>\n
\n
How\u00a0 Docker Model Runner works with Open WebUI<\/h2>\n
![\"Generate](\"https:\/\/www.docker.com\/app\/uploads\/2026\/03\/Generate-image-with-Open-WebUI-fig-1.png\" "\\"-")
\n <\/div>\nStep 1: Pull an Image Generation Model<\/h2>\n
\ndocker model pull stable-diffusion\n<\/pre>\n<\/div>\n
\ndocker model inspect stable-diffusion\n<\/pre>\n<\/div>\n
\n{\n \"id\": \"sha256:5f60862074a4c585126288d08555e5ad9ef65044bf490ff3a64855fc84d06823\",\n \"tags\": [\n \"docker.io\/ai\/stable-diffusion:latest\"\n ],\n \"created\": 1768470632,\n \"config\": {\n \"format\": \"diffusers\",\n \"architecture\": \"diffusers\",\n \"size\": \"6.94GB\",\n \"diffusers\": {\n \"dduf_file\": \"stable-diffusion-xl-base-1.0-FP16.dduf\",\n \"layout\": \"dduf\"\n }\n }\n}\n\n<\/pre>\n<\/div>\nStep 2: Launch Open WebUI<\/h2>\n
\ndocker model launch openwebui\n<\/pre>\n<\/div>\n
\ndocker run --rm \n -p 3000:8080 \n -e OPENAI_API_BASE=http:\/\/model-runner.docker.internal\/engines\/v1 \n -e OPENAI_BASE_URL=http:\/\/model-runner.docker.internal\/engines\/v1 \n -e OPENAI_API_KEY=sk-docker-model-runner \n ghcr.io\/open-webui\/open-webui:latest\n\n<\/pre>\n<\/div>\n
\ndocker model launch openwebui --detach\n<\/pre>\n<\/div>\n
Step 3: Configure Open WebUI for Image Generation<\/h2>\n
![\"Generate](\"https:\/\/www.docker.com\/app\/uploads\/2026\/03\/Generate-image-with-Open-WebUI-fig-2.png\" "\\"-")
\n <\/div>\n![\"Generate](\"https:\/\/www.docker.com\/app\/uploads\/2026\/03\/Generate-image-with-Open-WebUI-fig-3.png\" "\\"-")
\n <\/div>\n