Apr 3
Stream results
When dictating a large block of text or code, it is often helpful to be able to read the content as I go. For instance, I may need to refer back to something said/written earlier in order to refine my thoughts when constructing an email. With the current workflow I can achieve this by breaking it up into multiple superwhisper inputs, but this has a tendency to interrupt the flow of the thought process. It would be great if results could be either streamed or automatically broken into multiple inputs and continuously pasted into the text area. This way we aren't forced to either keep the entire content in our heads as we dictate or have to keep taking breaks to wait for processing to complete during an ongoing dictation.
Pending
This is now possible with a streaming model. A new model just came out that is able to handle streaming speech to text, which would enable you to show text as the user is speaking it instead of currently waiting for the user to complete. This is a potential game changer. Check out the model and try it here: https://kyutai.org/next/stt Would be great to have it as an option in Super Whisper!