You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
fix: resolve menu bar updates and AI enhancement issues
- Add immediate tray menu updates when model selection changes
- Fix default enhancement prompt to preserve natural speech format
- Add error feedback in recording pill when enhancement fails
- Implement event-driven API key removal with proper UI state sync
- Remove duplicate state updates and cache clearing calls
- Fix backend validation to allow empty provider for deselection
These changes ensure proper state synchronization across the app using
event-driven architecture, consistent with existing patterns in SettingsContext.
constDEFAULT_PROMPT:&str = r#"THEN clean up this voice transcription to create professional, readable text while preserving the speaker's intended meaning.
86
+
constDEFAULT_PROMPT:&str = r#"THEN clean up this voice transcription like a high-quality dictation tool would - fix errors while keeping the natural speech flow.
87
87
88
88
APPLY THESE CORRECTIONS:
89
89
90
90
1. REMOVE speech artifacts:
91
-
- Filler words: um, uh, ah, er, hmm, like (when filler), you know (when filler), basically (when redundant), actually (when redundant)
91
+
- Filler words: um, uh, ah, er, hmm, like (when filler), you know (when filler)
92
92
- False starts and incomplete thoughts
93
93
- Unintentional word repetitions and stutters: "I I think", "that that's"
94
94
- Verbal backspacing: "wait", "scratch that", "let me rephrase"
0 commit comments