What is the point of the desktop app?

[xPTM1219]

Probably a silly question, but honestly I don't see a good reason to use OpenWebUI Desktop app other than not wanting to use a browser. If I already have a server configured, why would I use the desktop app, when I can access through any browser.

I can be wrong, but I was trying the desktop app with the idea of setting up OpenAI and Ollama APIs for people without a server deployed. When I added Ollama's API, it simply loaded their webpage. I also added my local server, which loaded the login screen, and honestly at that point, I might go ahead and install the Docker version and completely ignore the desktop app.

If anyone has insight in having external APIs connected directly from OpenWebUI app and have the chat screen (not accessing a website), I would appreciate the instructions on how to do it.

Again, I can be wrong and was expecting a complete different thing, where I can connect directly to the APIs like is already possible to do in the normal deployment and chat with the different models. But this might not be the purpose of the app.

[laurikok]

If you install the llamacpp in the Interference Runtime tab of the OWU desktop and models in the model tab, you do not need a external API, which is partly the idea of the app. No external connections, everything local on the same machine in one configuration.

Other aspect is more fused experience of the front end with the spotlight, key shortcuts, screenshot based vision, maybe even wake word interference in the future. They are not possible or as integrated to the OS on a browser based frontend.