Problem

The current implementation in aws_services.py creates temporary audio file
in S3 but only deletes them after successful transcription. If transcription
fails or the process is interrupted, these files remain in S3 indefinitely,
which can:

• Lead to unnecessary storage costs
• Create potential security risks with stored audio files
• Violate data retention policies

Current Behavior

• Audio files are uploaded to 'audio-transcribe-temp' bucket
• Files are only deleted after successful transcription
• Failed transcriptions leave orphaned files
• No automated cleanup mechanism exists

Proposed Solution

1. Implement S3 bucket lifecycle policy to automatically delete objects afte
   24 hours
2. Add proper error handling to ensure cleanup in failure scenarios
3. Consider implementing a monitoring system for bucket usage

Technical Details

Affected files:

• aws_services.py
• Specifically the transcribe_audio() function

Implementation Notes

• Use AWS S3 lifecycle rules
• Consider adding CloudWatch metrics for bucket usage
• Add logging for cleanup operations