Feature Requests

right now, i can't seem to select my own mic. That's a showstopper for me to use the app. How could I select my own mic, which is different from my default mic?

Allow users to choose from different Whisper model options, including higher-quality models for transcription. This feature would cater to users with more powerful machines capable of handling larger models, providing them with the flexibility to prioritize transcription accuracy and quality based on their hardware capabilities.

I am running LM Studio with LLama3.3 70b (latest and greatest). I'd like to be able to use that to power summarization, chat, etc.

A workaround for lower CPU processing would be a way to simply record the audio, and then have the processing occur post meeting. Right now it is a live transcription; which I do not think is necessary in many cases. I think it would work well to allow me my full CPU power for the meeting, then run the transcription in the background post meeting and get my notes after some time.

It would be great when in a meetin to be able to identify speakers. Offer to isolate speaker voice in a small sample and they the user will associate the sample with the name of the participant. Ideally nutshell should remember this and reuse it for the next meeting.

This feature would allow users to upload audio files, such as recordings exported from MS Teams, into Nutshell for transcription. This enhancement will make the tool more versatile by enabling users to process recordings that were not directly captured through the platform. It should support commonly used audio formats (e.g., MP3, WAV) and maintain accuracy in transcribing diverse file types.

I set Nutshell as the speaker in Zoom. My own voice was captured (default Mic device), but not of the other speakers.