I'd also love if the OpenAI endpoint could be modified by the user! That'd make it very easy for us to pipe superwhisper's output to other local LLM services with OpenAI compatible endpoints, like Ollama or Local AI.

I have some beefy models running in Ollama I'd love to use with superwhisper, without duplicating the storage/RAM usage.

Neil. Is this picked up as part of your ealier request on API or separate. I’d like this option and it’d be amazing if it picks up my custom instructions like in GPT so that it knows my tone of voice and other details like where I work etc.