Open WebUI Notes

Installation and Set-up

NOTE: This assumes that you have downloaded Ollama beforehand. If not, please refer to the appropriate guide.

In a new terminal run (see below if this doesn't work):

docker run -it --name [INSERT CONTAINER NAME] -p 8080:8080 southerncrossai/modelworks:modelworks 

or, alternatively, to run container with GPU:

docker run -it --gpus all --runtime=nvidia --name [INSERT CONTAINER NAME] -p 8080:8080 southerncrossai/modelworks:modelworks /bin/bash

Use cd to move to the dir you want to work in, then run

python3 -m venv ./app/venv

source ./app/venv/bin/activate

pip install open-webui

After this is done, write:

open-webui serve

Then open a new terminal, and reopen your container:

docker exec -it "[INSERT CONTAINER NAME]" bash

And in your container write the following:

source ./app/venv/bin/activate

ollama serve

After this, you can open a browser and go to http://localhost:8080

---

Navigating Open WebUI

If this is your first time, you would need to create an admin account to access the rest of Open WebUI.

You can do so by entering your email, name, and any password you'd like.

Otherwise, just log in with your pre-existing credentials.

After this, you should be greeted by the following page:

If the appropriate instances are being run in the background, you can access your models around the top-left corner of the chat window:

• Again, this guide assumes you have downloaded Ollama in your container. If you haven't done so already, install the models you'd like to use.

Another important tab to visit is the Workspaces one located near the top-left corner of your whole screen:

From there you can locate your models, knowledge, prompts, and tools, which can help with customising and improving your UI:

Features

Response Ratings

When a model sends a response, Open WebUI gives the user the option to rate it:

This can be used for features involving assessing the confidence or satisfaction score related to the model's responses.

---