Queue: Spring Boot real-time transcription with Deepgram Java SDK

[github-actions]

Integration: Spring Boot + Deepgram Java SDK

Origin

Requested in #35: "A Spring Boot application that provides a WebSocket endpoint for real-time audio transcription using Deepgram's live STT API (Nova-3). Clients connect via WebSocket, send audio data, and receive live transcription results."

What this should show

• A Spring Boot 3.x application with a WebSocket endpoint that accepts audio from browser clients
• Proxies audio to Deepgram's live STT WebSocket API (Nova-3 model) via the deepgram-java-sdk
• Streams transcription results back to the browser in real time
• Includes a simple HTML/JS test page for browser-based microphone capture
• Uses Spring Boot conventions (application.properties for config, auto-configuration)
• Includes a health check endpoint
• Runnable with ./mvnw spring-boot:run after setting DEEPGRAM_API_KEY

Credentials likely needed

• DEEPGRAM_API_KEY

Reference

• Deepgram Java SDK: https://github.com/deepgram/deepgram-java-sdk
• Spring WebSocket docs: https://docs.spring.io/spring-framework/reference/web/websocket.html

---

Queued by PM from #35 on 2026-03-30