feat: Async recognition processing #3
Description
Description
Currently, voice recognition blocks until processing is complete. This creates a wait period where the user cannot do anything.
Problem
When using a push-to-talk hotkey, if you accidentally release and press again, everything recognized before gets erased. This is not ideal.
Proposed Solution
Process recognition asynchronously so users can continue working while audio is being transcribed.
Benefits
- No downside to async processing
- Prevents data loss from accidental hotkey release/re-press
- Allows seamless dictation without blocking workflow
- Users can continue working while longer passages are transcribed