feat: implement latest backend API integration for speaker diarization

Description

This MR adds frontend support for the latest backend Speaker Diarization API integration without introducing any backend changes.

The implementation enables users to toggle Speaker Diarization during transcription and ensures the correct diarization boolean is included in API requests based on user selection.

Changes Made

• Added Speaker Diarization toggle support in the frontend UI.

• Integrated latest backend transcription API contract changes.

• Updated transcription request payloads to include:

  diarization: true | false

• Updated related TypeScript interfaces and API typings.

• Added conditional handling for diarization-related API responses.

• Maintained compatibility with the existing transcription workflow.

• Improved state handling to avoid repeated API calls and render loops.

• Added proper cleanup and dependency handling for related useEffect hooks.

Screenshots

Testing

• Verified transcription works with diarization enabled.
• Verified transcription works with diarization disabled.
• Confirmed existing transcription flow remains unaffected.
• Tested API request payloads for correct diarization values.
• Validated no backend modifications were required.

Notes

• Backend APIs were not modified.
• Implementation follows existing frontend architecture and styling conventions.

closes #36 (closed)