Project: Plex Subtitle Orchestrator – Bazarr + OpenAI Whisper + ffsubsync

[lems111]

I’ve built a small companion stack around Bazarr that might be useful to others here.
The idea:

• Let Bazarr do what it’s good at: normal subtitle downloads.
• When a show/movie has no subs, run an on‑demand Whisper container to:
  • extract audio,
  • call OpenAI Whisper for transcription to ${LANG}.srt,
  • translate to .en.srt using GPT.
• Run a separate ffsubsync container that watches the media tree and automatically syncs .en.srt files whenever they appear.

It’s all Dockerized with:

• Compose file defining Bazarr + subsync + whisper-runner
• Custom images that bake in ffsubsync / Whisper dependencies (no per‑run apt/apk setup)
• Example .env and README with Bazarr config notes (e.g., leaving Bazarr’s own auto‑sync off)

Repo: https://github.com/lems111/plex-subtitle-orchestrator

• I’d be very interested in feedback from Bazarr/Whisper users, especially:
• Better ways to integrate with Bazarr’s workflows
• Any gotchas you’ve hit with subsync/Whisper in similar setups
• Suggestions for improvements before I call this a “1.0”

Happy to adjust things if there are patterns you’d recommend instead.