OpenAI Brings GPT-Live Voice to Desktop Apps

OpenAI documents three desktop experiences with distinct scopes:

• •Chat handles conversational assistance, search, brainstorming, and everyday questions.
• •Work is intended for longer, multi-step work such as research and creating documents, spreadsheets, presentations, reports, or sites.
• •Codex is dedicated to software development tasks, including writing or debugging code, running tests and commands, reviewing changes, and working with repositories.

The company documentation states that Voice can start separate threads for longer tasks, inspect existing threads, and relay progress, blockers, and results back into the voice conversation. An example in the documentation is asking Codex to run tests and investigate failures, then checking active tasks for blockers.

Android Authority reports that the desktop rollout is global and cites an OpenAI post describing Voice as a way to direct multiple agents in Work or Codex. That post states: "Control your computer and direct multiple agents running in ChatGPT Work or Codex, using just your voice."

On macOS, users can enable Screen context in Voice settings and ask ChatGPT to inspect the frontmost window. OpenAI calls the resulting capture an AppShot. Its documentation states that an AppShot may contain the window image and accessible text, and that organizations can disable the capability.

This permission model matters for enterprise deployments because screen captures and accessible text can expose information from a user's active application. OpenAI's documentation indicates that workspace settings affect availability, while the organization can disable screen context. Teams using regulated or sensitive desktop workflows would therefore need to evaluate the applicable workspace controls and data-handling configuration before enabling the feature.

ChatGPT Voice is available in the desktop app for Plus, Pro, Business, Edu, and Enterprise plans, according to OpenAI. Enterprise and Edu access begins with a two-week early-access period before default availability, subject to rollout status and workspace settings. OpenAI also notes that users can access Voice after pairing a phone with a desktop host.

For developers, the notable change is not voice dictation alone but a voice interface connected to task-oriented Work and code-oriented Codex workflows. Comparable agent interfaces often make task state, permission boundaries, and interruption behavior central operational concerns. OpenAI's documentation frames its implementation around voice-directed task coordination while preserving the tools and permissions of the selected Chat, Work, or Codex experience.