What is Open WebUI?
Open WebUI is a self-hosted chat interface for local models. It looks and works much like ChatGPT, keeps conversation history, supports several users with their own accounts, and can answer questions about uploaded documents. It talks to Ollama for the models themselves.
What do you need first?
- Docker or the vendor's container manager on your NAS
- Ollama running on the NAS (see our Ollama guide), or the bundled image in step 2
- Enough memory for your chosen model plus around 1GB for Open WebUI
Step 1: Choose the right image
The project publishes two images. ghcr.io/open-webui/open-webui:main is the interface on its own and connects to an existing Ollama. ghcr.io/open-webui/open-webui:ollama bundles both in one container.
Step 2: Create the container
With Ollama already on the same NAS, the official command from the Open WebUI README is:
`docker run -d -p 3000:8080 --add-host=host.docker.internal:host-gateway -v open-webui:/app/backend/data --name open-webui --restart always ghcr.io/open-webui/open-webui:main`
If Ollama runs on a different machine, replace the add-host option with `-e OLLAMA_BASE_URL=` followed by that machine's address. For the bundled image, map the ollama volume as well and use the :ollama tag.
Step 3: Open it and create the admin account
Browse to port 3000 on your NAS address. The first account you create becomes the administrator. Add accounts for family or colleagues from the admin panel, and turn off open sign-up if the NAS is reachable from outside your home.
Step 4: Pick a model and start chatting
Open WebUI lists the models Ollama has downloaded. You can pull new ones from the settings page. Upload documents to a chat to ask questions about them.
How do you keep it private?
Leave cloud API connections switched off so every request stays on the NAS, keep the port off the open internet, and use the vendor's VPN or secure remote access when away from home.
Commands from the Open WebUI README.
Common questions
Does Open WebUI need a GPU?
No. It is a web application. The speed of answers depends on the hardware running the model.
Can several people use it at once?
Yes. Each person has their own account and history, sharing the same models.
How do I update Open WebUI?
Pull the latest image and recreate the container with the same data volume. Chats, users and settings live in that volume, so they carry over. Read the release notes before major updates, as settings occasionally move.
Can Open WebUI search my documents?
Yes. You can attach files to a chat or build knowledge collections that the model searches before answering. For a larger document library, see our guide to private RAG on an AI NAS.
Can I reach Open WebUI from outside my home?
Use a VPN such as WireGuard or Tailscale, or the NAS maker's own secure remote access, rather than opening a port to the internet. That keeps the login page off the public internet while still letting you chat from your phone.
