Use ollama with local llm

[M0dM]

Hi,

Would you think it would be relevant to test using ollama ?

auto-inspector could support multiple LLM backends and pass ollama "backend url" like OPENAI_API_KEY value is actually set.

I would be very interested to find out if it is giving good results using a simple model like codellama-7b.

Ref:
backend/src/infra/services/openai4o.ts