Queue: Django Channels real-time transcription with Deepgram Live STT #65
Description
Integration: Django Channels + Deepgram Live STT
Origin
Requested in #34: "A Django application using Django Channels for WebSocket support that captures browser microphone audio and streams it to Deepgram's live STT API in real-time, displaying transcription results on the page."
What this should show
Build a Django 5.x application using Django Channels that:
- Captures browser microphone audio via a simple HTML/JS page (no separate frontend build)
- Streams audio over a Django Channels WebSocket consumer to the server
- Server-side consumer forwards audio to Deepgram's Live STT WebSocket API (Nova-3 model)
- Transcription results are sent back to the browser and displayed in real-time
- API key stays server-side only
- Uses ASGI consumer pattern with Django Channels
- Python 3.11+, current deepgram-sdk version
Credentials likely needed
- DEEPGRAM_API_KEY
Reference
Original issue has detailed acceptance criteria and architecture description.
Queued by PM from #34 on 2026-03-30