[Suggestion] Spring Boot voice agent with Deepgram Java SDK (Java)

[deepgram-robot]

What to build

A Spring Boot application that implements a real-time voice agent using Deepgram's Voice Agent API and the new Deepgram Java SDK, targeting enterprise Java developers and teams migrating from discontinued competitor SDKs.

Why this matters

Spring Boot is the dominant Java application framework in enterprise environments. With the Deepgram Java SDK recently launching (v0.1.0, March 2026), Java developers need a production-quality reference implementation showing how to build voice AI applications in their ecosystem. Many enterprise Java teams are evaluating voice agent capabilities, and a Spring Boot integration provides the idiomatic patterns they expect — dependency injection, configuration properties, WebSocket support via Spring WebSocket, and production-ready error handling.

Suggested scope

• Language: Java 17+
• Framework: Spring Boot 3.x with Spring WebSocket
• Deepgram APIs: Voice Agent API (WebSocket), STT (Nova-3), TTS (Aura)
• What it should include:
  • Spring Boot auto-configuration for Deepgram client
  • WebSocket handler for browser ↔ Spring ↔ Deepgram Agent API
  • Configuration via application.yml (API key, model selection, voice)
  • Simple Thymeleaf or static HTML frontend for voice interaction
  • At least one function calling tool demonstrating enterprise integration
  • Gradle and Maven build files
• Complexity: Medium

Acceptance criteria

• Runnable with minimal setup (clone, add API key, ./gradlew bootRun)
• README explains the Spring Boot architecture clearly
• Uses current Deepgram Java SDK version
• Demonstrates voice agent with function calling
• Follows Spring Boot conventions (auto-config, properties, component scanning)

---

Raised by the DX intelligence system.