OpenAI brought its GPT-Live voice mode to the ChatGPT desktop app on July 23, 2026 — a spoken interface that lets you control your computer and direct multiple agents in ChatGPT Work and Codex while it speaks, listens, and coordinates at the same time.
2026-07-24 · by Moe Ameen
OpenAI rolled ChatGPT Voice into the desktop app on July 23, 2026, bringing a spoken interface to macOS and Windows for Plus, Pro, Business, Edu, and Enterprise plans. It runs on GPT-Live, the voice-model family OpenAI introduced earlier in July 2026 on mobile, and OpenAI pitches it as its most advanced voice experience yet: instead of a back-and-forth chat, you talk through what you need and ChatGPT starts moving several tasks forward at once. You launch it with a programmable hotkey or a Voice button and start a chat or task in voice mode.
The headline capability is agent control by voice. OpenAI frames it as being able to "control your computer and direct multiple agents running in ChatGPT Work or Codex, using just your voice" — the model can speak, listen, and coordinate work in the app simultaneously. A launch demo showed a developer issuing one spoken command to create a thread, open a pull request, and hunt down the root cause of a bug. It works alongside OpenAI's Computer Use feature, local files, and ChatGPT plugins, and on macOS a feature called Appshots lets the voice mode reference the active window on screen for context. Voice can also drive Codex from the iOS app through the Remote feature by pairing to a computer, with Android support described as coming.
Positioning-wise this is aimed squarely at developers and AI power users — talking through code, planning a trip by pulling from your calendar and email, or dictating a document while thinking out loud. It brings the GPT-Live voice experience that launched on mobile to the macOS and Windows desktop apps. Treat the specifics here as a launch-window snapshot; rollout is gradual and OpenAI's own pages are the place to confirm current plan availability and platform support.
The interesting thing for creators is the motion ChatGPT Voice unlocks: you can now think out loud into a finished draft. Walk through your next post idea, dictate the angle, argue with yourself about the hook — and get clean text back without touching the keyboard. That is a real upgrade to the front of the workflow. But it stops exactly where every ideation tool stops. GPT-Live hands you one spoken-into-existence draft in one window. It does not fan that idea across formats, and it does not publish. Kompozy is the operation that turns a single idea into a shipped content week.
Use the two together and the seam disappears. Talk your idea into shape in ChatGPT Voice, then drop that spark into Kompozy and let it generate the full set in one voice through your Persona Brief — a Persona Short with auto-captions, a brand-exact Carousel, a Quote Graphic, an X thread, a Blog Article, and an Email Newsletter — then auto-caption, reframe, and publish the batch across the eight social platforms plus blog and email from one review queue. And if "talk through what you need and it starts moving tasks forward" is the ethos you like, Kompozy's autopilot is the version that runs on your content: point it at your sources, set the cadence, and it generates, schedules, and publishes on a recurring basis without you narrating each step. ChatGPT Voice makes ideation conversational; Kompozy makes the idea a published, on-brand operation across every platform.
It is a spoken interface added to the ChatGPT desktop app on July 23, 2026, running on OpenAI's GPT-Live voice models. You launch it with a hotkey or a Voice button and talk through what you need; it can control your computer and direct multiple agents in ChatGPT Work or Codex while it speaks, listens, and coordinates work at the same time.
At launch it rolled out on macOS and Windows for Plus, Pro, Business, Edu, and Enterprise plans, gradually. On macOS it adds Appshots, which lets voice mode reference the active window for context. It also works with Computer Use, local files, and plugins, and can drive Codex from the iOS app through the Remote feature. Confirm current availability on OpenAI's own pages, since rollout is phased.
It can help you draft by voice — dictate an idea, a script, or a document hands-free — but it produces one output in one window and does not publish anything. Reframing an idea into a short, a carousel, a thread, a blog, and a newsletter, then scheduling across platforms, is not part of it.
Use ChatGPT Voice to talk an idea into a clean draft, then bring that spark into Kompozy. Kompozy generates the full format set in one brand voice — persona/avatar video, carousels, quote graphics, threads, blogs, and newsletters — auto-captions and reframes each, and publishes across the eight social platforms plus blog and email from one queue, or runs the whole cadence on autopilot.