There are Base URL for OpenAI only in Base Model sub-node. Why? Somewhere it was said to prevent issues as there are lots of differencies on different OpenAI Based services such as XInference for example.
I could understand that if n8n was non-dev friendly but it is not really true.
Also I can not understand why I can use Base URL in OpenAI Model sub-node but cannot use it in such node as Message a model.
1 Like
Hey @mr-good-bye,
I think it was in the older version of the OpenAI node, But as the OpenAI features got more complex a lot of the “compatible” services didn’t function as well with the node so it was safer to move to the model specific options.
At some point we did think about making a generic OpenAI node where you can specify the URL but at the moment it isn’t a priority.
You can still message a model by using the model node with the AI Agent which should do the same thing for now.
I ran into the same issue and believe that having this capability is important, especially as N8N expands to support more frameworks.
For example, I wanted to use the API provided by LM Studio in N8N, but I couldn’t do so because the OpenAI model node doesn’t allow setting a base URL.
Below, you can see that their documentation mentions they use the OpenAI API structure.
