TTS fails with uncaught errors for unauthenticated users

[rosscado]

Problem

When an unauthenticated user uses SayPi, the extension attempts to synthesize speech (TTS) which fails because TTS is a premium feature. This results in repeated uncaught promise rejections in the console:

Uncaught (in promise) Error: Failed to synthesize speech
    at TextToSpeechService.createSpeech
    at async AudioStreamManager.createStream
    at async SpeechSynthesisModule.createSpeechStream
    ...

Expected Behavior

1. Voice selector should be disabled/greyed out when user is unauthenticated (to indicate TTS is unavailable)
2. Don't attempt TTS synthesis when it will predictably fail
3. Handle errors gracefully - no uncaught promise rejections in console

Current Behavior

• Voice selector is enabled and selectable even when unauthenticated
• Extension attempts TTS on every assistant message, failing each time
• Errors are unhandled, cluttering the console

Suggested Fix

• Check auth status before attempting TTS
• Disable voice UI elements when TTS is unavailable
• Wrap TTS calls in proper try/catch with graceful degradation

🤖 Generated with Claude Code