OpenAI Brings Full-Duplex Voice Control to Codex and ChatGPT Desktop

Executive Briefing

• Enables hands-free software development by integrating GPT-Live voice into Codex and ChatGPT Work desktop environments on macOS and Windows
• Supports simultaneous multi-task execution, letting developers verbally trigger bug fixes, pull request reviews, and unit test generation concurrently
• Decouples real-time voice layer from background reasoning models, maintaining fluid conversation while delegating heavy computation to GPT-5.5
• Reaches over 10 million weekly active Codex and ChatGPT Work users, marking the first voice activation integration for these platforms
• Restricts access to paid subscribers only; model weights and pipelines remain fully closed with no self-hosting option