Skip to main content
Back

OpenAI Ships GPT-Live Voice to Desktop, Turning Speech Into Agent Control

ChatGPT Voice lands on macOS and Windows desktops, letting you control your computer and direct multiple AI agents using only your voice.

  • ChatGPT Voice is now live on macOS and Windows for Plus, Pro, Business, Edu, and Enterprise plans, rolling out globally today.
  • Powered by GPT-Live, a full-duplex voice model that listens and speaks simultaneously, unlike the old turn-based Advanced Voice Mode.
  • Voice controls the entire desktop app, including computer use via Codex and multi-agent orchestration in ChatGPT Work.
  • Two-layer architecture: a lightweight voice model handles conversation flow while GPT-5.5 handles complex reasoning and search in the background, seamlessly.
  • Part of a broader consolidation: Codex, Chat, and Work are now unified in one ChatGPT desktop app, with voice as the top-level control layer.
  • Security surface is larger: voice-triggered computer use on real machines warrants careful review before broad enterprise deployment.

OpenAI has shipped ChatGPT Voice to the desktop app, rolling it out globally on macOS and Windows for Plus, Pro, Business, Edu, and Enterprise plans. The headline feature is voice-controlled computer use and multi-agent orchestration, all driven by a single spoken command.

The engine underneath: GPT-Live

The feature runs on GPT-Live, the voice model architecture OpenAI launched on July 8, and it represents a genuine architectural departure from what came before.

Previous voice systems were half-duplex: the model waited for you to stop talking, then responded. Advanced Voice Mode processed and generated audio within a single model but still operated through discrete turns, meaning even a brief pause or background noise could be mistaken for the end of a turn.

GPT-Live runs on a full-duplex architecture that processes input and output simultaneously. It is also a two-layer system: a continuous interaction layer manages the conversation, while a delegation layer quietly hands complex tasks to GPT-5.5 in the background. The system makes decisions multiple times per second about whether to speak, listen, pause, interrupt, or invoke tools, and incorporates natural cues like "mhmm" and "got it" to signal engagement. The result feels less like querying a model and more like talking to a colleague who is already pulling up the file while you finish your sentence.

What you can actually do with it

Voice is a control layer for the entire ChatGPT desktop app, which now includes Codex alongside Chat and Work.

  • Computer control: Speak to control your machine directly. Codex can see, click, and type in Windows applications, including specific apps invoked from prompts.
  • Multi-agent orchestration: Direct multiple agents running in ChatGPT Work or Codex simultaneously using only your voice.
  • Background task delegation: When a question requires web search, deeper reasoning, or agentic capabilities, GPT-Live sends the task to GPT-5.5 in the background and keeps the conversation going. When GPT-5.5 finishes, GPT-Live folds the result naturally into the conversation with no gap or pause.
  • Parallel task management: The voice layer maintains conversation while worker agents handle tasks in the background. Documented examples include checking Slack, pulling up documents and calendars, controlling Spotify, browsing, shopping, and food ordering.

One app, everything in it

This release caps a consolidation strategy OpenAI has been executing for months. Codex was merged into a revamped ChatGPT desktop app with a built-in browser and computer control. The old ChatGPT desktop app has been renamed ChatGPT Classic, and users with the Codex app installed can update it to get the new unified version.

Adding voice on top of this unified app turns voice into an orchestration layer rather than a chat modality. Dispatching and managing several tasks simultaneously on a laptop, with results read back to you, is a fundamentally different use case than asking ChatGPT a question out loud.

Where it fits in real workflows

The clearest wins are in hands-free, high-context situations:

  • Coding sessions: Describe a bug verbally while Codex reads your screen and starts iterating, without switching context to type a prompt.
  • Research and writing: Ask ChatGPT Work to pull sources, draft a section, and open relevant documents while you talk through your outline.
  • Meeting prep: Speak a request to check your calendar, pull a Slack thread, and summarize it while you are already walking to the room.
  • Accessibility: For users who find keyboard-heavy workflows difficult, voice control over a full agentic desktop environment is a substantial quality-of-life improvement.

Limits and risks worth knowing

GPT-Live-1 is the default for Go, Plus, and Pro users; GPT-Live-1 mini is the default for Free users. The desktop voice feature requires Plus, Pro, Business, Edu, or Enterprise. Free users are excluded from this rollout.

The expanded surface area carries real security implications. OpenAI's own documentation warns that Codex's broader capabilities increase exposure to prompt injection attacks, where malicious instructions embedded in a webpage can enter the memory store and later be executed by Codex. Giving voice commands that trigger computer-use agents on a machine with real files is powerful, and worth thinking through carefully before deploying in sensitive environments.

Google's Gemini Live and Meta's voice-enabled assistant have both pursued more natural, real-time conversation, but neither is currently wired into a full agentic desktop environment. Voice as an agent orchestration layer is the specific bet OpenAI is making here, and for now it has no direct equivalent from either competitor.