Use LocalAI in paperless-ai

[oswaldo-be]

Hello,
I am trying to use a localai model for my paperless-ai.
I tried it with llama and gpt-4o.
Can anybody explain how it can be integrated.
"Error
Custom connection failed. Please check URL, API Key and Model."
Did anyone set it up like this and can give me a hint?
Thank you very much

[clusterzx]

Does it feature the OpenAI API structure?

[oswaldo-be]

It seems so
https://localai.io/features/openai-functions/

[clusterzx]

@oswaldo-be ha e you tried to check your network settings?

Use host.docker.internal instead of localhost. Maybe that helps already.

[oswaldo-be]

This is the way from any Browser at home to chat with this assistant.
http://192.168.178.60:8080/chat/llama-3.2-3b-reasoning-time
I also created an "API-key" but also without I see this error.

[clusterzx]

Yeah but the url you posted is clearly not the API address of LocalAI

[clusterzx]

Th correct api url to enter would be http://192.168.178.60:8080/v1/

[oswaldo-be]

Oh my. Thank you very much.
Of course it is just the IP and the port and in the field name I just have to add the name of the model....

[clusterzx]

Ok if everything is ok you can mark my comment as answer and the post will be closed.

If not feel free to post here again.