1. Update UI to show actual _currentText instead of fixed 'realtime recognizing...' text - Use Flexible widget to handle text overflow - Show actual recognition text when available - Fallback to 'realtime recognizing...' when text is empty 2. Fix final result processing logic - Don't update _currentText when processing final results - Final results only append to text field, don't interfere with real-time display 3. Improve listening status management - Clear _currentText when recording stops - Ensure clean state transitions between recording sessions This fixes the issue where users could only see final results without real-time feedback. |
||
|---|---|---|
| .. | ||
| main.dart | ||