AI setup with local models

[royalj7]

Hey folks,

I was hoping someone could help me with getting the AI features enabled in Calibre with my setup. My setup is I run Ollama and OpenWeb UI on a dedicated box and interact with local models using OpenWeb UI's GUI. Ollama is installed as a program and OpenWeb UI is run in docker using Docker Desktop. Calibre is run on my NAS that is a separate machine. The key is I don't expose Ollama to the network, only the localhost. OpenWeb UI is exposed across the network. In Calibre's settings I've tried OpenAI compatible for the AI provider and also OllamaAI and using OpenWeb UI's IP address and port and have not been successful.

Anyone have any experience with this type of setup?

Appreciate the help,
--Jamie