Addresses #365

/claim #365

Completed Objectives

This PR: achieves the following:

• Extend existing OpenAI Whisper implementation to handle longer audio files
• Implement audio chunking for files over 20 minutes
• Use Vercel Blob for storing audio files
• Update plugin to use new transcription endpoint
• Remove the need for the dedicated audio-server

Limitations

• Due to the inability to get the plugin UI up and working in a local development setup:
  • Did not integrate the Backend Server to the UI plugin.

Short audio Transcription

https://github.com/user-attachments/assets/cb5cf0e3-862e-4ef2-80d1-4c8636167b1e

50 min audio transcription

Chunk the audio file to < 20 min and <25 mb

final result