Basic idea
Open WebUI is a self-hosted chat application that you can run entirely on your DGX Spark. This gives you privacy and security because everything stays on your system. The Open WebUI application operates offline, and your queries go to a model running on your Spark.
Open WebUI is browser based, so you can use it to chat with a model on a Spark from your laptop, or directly on the Spark itself as a desktop.
This playbook shows you two different paths to run Open WebUI.
- Remotely with NVIDIA Sync: Use this path to run Open WebUI on a remote Spark from your laptop.
- Manually on a Desktop: Use this path to run Open WebUI on a desktop Spark or if you want to get under the hood.
Both paths lead to the same outcome: chatting with a model running on your DGX Spark through Open WebUI.
What you'll accomplish
You will download an Open WebUI container image with Ollama onto your Spark, run the container and then use the Open WebUI browser interface to download and run a model. Then, you will chat with it.
The container setup includes integrated Ollama for model management, persistent data storage, and GPU acceleration for model inference.
What to know before starting
- The Ollama models page to get information on models, e.g. qwen3.6 and gpt-oss
- Remote with NVIDIA Sync: Cut and paste terminal commands; How to use NVIDIA Sync. See documentation here and installation here.
- Manually on a Desktop: Terminal experience; Familiarity with Docker commands
Prerequisites
- DGX Spark device is set up and accessible
- Enough disk space for the container image and models (7GB for container image, 25GB for qwen3.6
or 15GB for gpt-oss ) - Remotely with NVIDIA Sync path: NVIDIA Sync installed on your laptop and connected to the Spark to your DGX Spark
Time & risk
- Duration: 15-20 minutes for setup, includes Open WebUI container download as well as model download (time varies per your internet speed)
- Risks:
- Docker permission issues may require user group changes and session restart
- Large model downloads may take significant time depending on network speed
- Last Updated: 07/25/2026 (minor copy edits)