implement prompt post-processing option

[swiftraccoon]

https://platform.openai.com/docs/guides/speech-to-text/improving-reliability

[swiftraccoon]

could see this being extra helpful if we implement #13 . design our own local dictionary corrections?

[swiftraccoon]

OpenAI GPT4 API is still too expensive, but now that we've got local transcription setup regarding #13 maybe now it is time to investigate implementing a local dictionary with local LLM and see if it helps any.