Summary

I refactored the aria-live announcements to improve compatibility with macOS VoiceOver, enhancing the accessibility of LibreChat for screen reader users.

• Simplified the LiveAnnouncer component by removing chunking and queueing mechanisms.
• Separated live regions into "status" and "log" areas for clearer announcements.
• Implemented a status region for short, process-related announcements.
• Created a log region for complete messages.
• Introduced a simple delay between status updates to prevent overriding.
• Added periodic status updates for long-running AI responses (~7 seconds)
• Updated localization strings to support new announcement types.
• Modified the SSE event handlers to use the new announcement system.

Change Type

• New feature (non-breaking change which adds functionality)
• This change requires a documentation update

Testing

I tested the changes with both NVDA and VoiceOver screen readers. The new implementation provides a more consistent and interruption-free experience, especially for macOS VoiceOver users.

Test Configuration:

• Test with NVDA on Windows
• Test with VoiceOver on macOS
• Verify announcements for different chat scenarios (short responses, long responses, streaming responses)

Checklist

• My code adheres to this project's style guidelines
• I have performed a self-review of my own code
• I have commented in any complex areas of my code
• I have made pertinent documentation changes
• My changes do not introduce new warnings
• Local unit tests pass with my changes