Deepgram Voice Agent APIdeepgram-sdk (Python/JS)WebSocketsOpenAI (LLM)Web Audio API (browser)intermediate30 min

Deepgram + real-time speech agent

Real-time bidirectional voice conversations with natural turn-taking, barge-in handling, and LLM orchestration for AI agents.

Prerequisites

  • Deepgram API key
  • Python/Node.js/React env
  • deepgram-sdk (`pip install deepgram-sdk`)
  • OpenAI API key for think provider
  • microphone access
  • supported audio formats (linear16 PCM preferred).

Further reading