GPT-Live-1 Hits the API: Build Real-Time Voice Agents with Full-Duplex Audio and Tool Calling

Tibo ·

Key Info

OpenAI has released GPT-Live-1 in the API, letting developers build voice agents with the same full-duplex, natural back-and-forth voice system used by ChatGPT's 1 billion users.

Highlights

  • Full-duplex conversation: the model listens while it speaks, enabling natural interruptions and real-time back-and-forth.
  • Supports tool calling, so voice agents can take actions as well as converse.
  • Flexible integration: developers can pair it with the models and speech harness of their choice.
  • Built on the voice system already shipped to 1B ChatGPT users.
Loading...