Add configurable timeout for OpenAI-compatible providers

[lyesrock]

Hi,
Could you please add the ability to adjust the request timeout for all OpenAI-compatible endpoints, not just Ollama?

We run large local models (llama.cpp, LM Studio, Ollama OpenAI mode, etc.) with big context windows, and we hit the same timeout issues. Since these setups behave like Ollama in terms of latency, having the “Request Timeout (ms)” option available for OpenAI-compatible models would solve the problem.

Thanks!

[thor171]

I would appreciate that too. For local usage, it is a neccessary feature, since some requests take a long time.