OpenAI updated its ChatGPT desktop app on July 24, 2026, to include ChatGPT Voice, allowing users to control AI agents and perform multi-step tasks on their computers using spoken commands.
The feature is powered by ChatGPT-Live, a family of voice models OpenAI launched earlier in July 2026. It works with both ChatGPT Work and Codex, and can access websites and apps through computer use capabilities. On macOS, a feature called Appshots lets the app read what is on the user’s screen, including alt-text.
The desktop version is more capable than the existing smartphone version of ChatGPT Voice, which launched with smoother conversations and better interruption handling but was not designed to take action on devices. The desktop update allows users to issue complex, multi-step commands and respond when ChatGPT requires additional input during a task.
In a demo, OpenAI showed a developer using a single voice command to ask ChatGPT to create a new thread, submit a pull request, and identify the root cause of a bug — illustrating how the feature may reduce the number of manual steps required for technical workflows. Users can also access ChatGPT Voice in Codex remotely through the iOS app.
The update is rolling out globally. Separately, Anthropic has also updated its voice mode for Claude, which can use the company’s Opus, Sonnet, and Haiku models to complete tasks in apps including Gmail, Calendar, Slack, Notion, and Canva.
Source: TechCrunch