Feature request: send context around cursor as prompt

[heimoshuiyu]

An important feature of Whisper is its ability to use specific nouns from the context to improve speech recognition accuracy.

So I was thinking that when I write some proper nouns in my notes and start voice input, the text around my cursor can be sent to the API as prompts. This way, Whisper will know the context I am in and the keywords I am using during transcription, which can greatly improve the recognition rate. ❤️

To be specific , I believe there are several possible solutions that can be implemented:

1. Send the entire document opened by the user as a prompt to the API.
2. Send the text above the cursor as a prompt to the API.
3. Send only the text highlighted by the user as a prompt to the API.

For the last solution, I'm not sure if this can be achieved through a plugin and whether the same effect can be accomplished on a mobile application.