Description

Integrate LLM models using LLM Studio with Open WebUI

Deployment

Follow the next steps to integrate LLM Studio with Open WebUI

• STEP01: Download LLM Studio and install in Ubuntu 24.02. Download LLM Studio for your architecture. In this case you will download a AppImage file. Copy this file under /opt/LMStudio, set some priviledges:

$ chown own simur:simur /opt/LMStudio
$ chmod +x /opt/LMStudio/LM_Studio.AppImage

Also download a LLM Studio icon lmstudio.png inside this repo under the same AppImage file folder.

• STEP02: Create a file launcher called llm-studio.desktop under ~/.local/share/applications like this:

[Desktop Entry]
Name=LM Studio
Comment=LM Studio Launcher
Exec=/opt/LMStudio/LM_Studio.AppImage --no-sandbox
Icon=/opt/LMStudio/lmstudio.png
Terminal=false
Type=Application
Categories=Development;
StartupWMClass=LM Studio

• STEP03: Install a LLM Model and Load Now we can start LLM Studio, and download our first LLM model to be used. Click in LM Runtime localed at left top and select a model to be downloaded and later loaded. In this case we select the model called openai/gpt-oss-20b).

After be download we can load it. If we go to Developer menu option we see our model loaded

• STEP04: Start LLM Server and configure

Now with a LLM Model loaded, we can start our LLM Server. LLM Studio start a App icon menu at top of ubuntu.

Now we configure this new LLM Server from LLM Studio. Go again to Developer LLM Studio and click Server Settings and set this options:

With these options we can accedd to the LLM Server with port 1234 externally

• STEP05: Install Open WebUI We will install Open WebUI from a docker container like this:

$ docker run -d -p 3000:8080 -v open-webui:/app/backend/data --name open-webui ghcr.io/open-webui/open-webui:main

• STEP06: Configure Open WebUI Now we must create a connection with out LLM Server from Open WebUI. Open WebUI from your browser: http://localhost:3000 and register a first account in it. Later go to Settings under your account

And inside settings select Admin Settings at botton. Select the option Connections to create our OpenAI connection with LLM Studio Server

Click under plus icon and set this url: http://<LLM_STUDIO_SERER_HOST>:<LLM_STUDIO_SERER_PORT>/v1 without any authentication (set Auth to None)

• STEP07: Use your models Now we can select any model deployed in my LLM Server and send a prompt request like this: