Skip to main content

ChatGPT Desktop Gets Voice Control for Multi-Step Tasks

OpenAI is taking desktop AI interaction to the next level. On Thursday, the company announced a significant update to its ChatGPT desktop application, introducing full voice control. Now, instead of typing, you can simply talk to ChatGPT and have it carry out complex, multi-step tasks on your computer—hands-free.

This new capability is powered by the ChatGPT-Live series of speech models, which OpenAI rolled out earlier this month. The voice feature works seamlessly with ChatGPT Work and Codex, and it's built to handle real-world computer operations. On macOS, it even integrates with Appshots, letting you authorize the AI to see everything on your screen, including alternative text for images.

What's really exciting is how far this goes beyond the mobile voice features we've seen before. Those were mostly about smooth conversation and handling interruptions. This desktop version is a whole different beast—it's about getting things done. You can give compound commands in plain language, and if ChatGPT needs more info or wants to double-check something, it'll ask you interactively.

In a demo, a developer used a single voice command to create a new code thread, submit a pull request, and quickly pinpoint the root cause of a bug. That's the kind of workflow that used to require a keyboard, a mouse, and a lot of clicking. Now it's just a matter of speaking.

OpenAI also mentioned that you can use the remote access feature of the iOS app to trigger ChatGPT Voice in Codex, which adds even more flexibility.

But OpenAI isn't the only player in this space. Anthropic has also been busy. They've introduced a new voice mode for Claude, their AI assistant. This mode can tap into Claude's Opus, Sonnet, and Haiku models to help with everyday tasks in popular office tools like Gmail, Calendar, Slack, Notion, and Canva.

The race to make AI more conversational and more capable is heating up. With voice control becoming a standard feature, the way we interact with our computers is about to change dramatically. Whether you're a developer, a busy professional, or just someone who likes to talk to their tech, these updates are worth paying attention to.

Key Points

  • Voice control on desktop: ChatGPT's desktop app now supports voice commands for executing multi-step tasks.
  • Powered by ChatGPT-Live: The feature uses the latest speech models from OpenAI.
  • Integration with Work and Codex: Works with ChatGPT Work and Codex, and on macOS with Appshots for screen access.
  • Real-world demo: A developer used voice to create a code thread, submit a PR, and find a bug.
  • Competition: Anthropic's Claude also introduced a voice mode for office applications.