feat(transcripts): add audio transcript generation and editing workflow

[eddmann]

Summary

Add transcript generation for podcast recordings with multi-voice support and a workflow for editing/cleaning transcripts.

Description

We want to generate and provide transcripts with each recording. This involves:

1. Transcript Generation: Review options for automatic transcript generation

   • Paraqeet's multi-voice options look interesting for handling multiple speakers

2. Transcript Editing Workflow: Need a means to:

   • Go through the audio alongside the transcript
   • Clean up and edit the transcript as we go
   • Produce a final transcript in a sensible format

Requirements

• Research transcript generation options (including Paraqeet multi-voice)
• Implement transcript generation for recordings
• Create an editing interface to review audio with transcript
• Support cleaning up/editing transcripts while listening
• Define and implement a sensible output format for final transcripts

Additional Context

This is for a podcast workflow where accurate speaker attribution and clean transcripts are important.