OpenAI Brings ChatGPT Voice to Desktop With AI Agent Controls

OpenAI has expanded its ChatGPT desktop application with a new voice mode that enables users to speak directly to the AI to control agents and perform tasks on their computers. Announced on Thursday, the feature brings the company’s recently introduced ChatGPT-Live voice models to desktop users, making voice interactions more capable than those currently available on smartphones.

The new ChatGPT Voice feature works with both ChatGPT Work and Codex, allowing users to issue complex spoken commands involving multiple steps. It can also use computer interaction capabilities to access websites and applications, while macOS users can enable Appshots to let ChatGPT view on-screen content, including alt-text, when needed.

Voice commands for desktop tasks.

Unlike the mobile version, which focuses on natural conversations and improved interruption handling, the desktop experience is designed to complete real tasks. Users can dictate multi-step instructions, and ChatGPT can pause to request additional input before finishing an assignment.

In a demonstration, OpenAI showed a developer asking ChatGPT to create a new thread, open a pull request and identify the root cause of a software bug using a single voice command. The company also said users can access ChatGPT Voice in Codex remotely through the ChatGPT iOS app.

Competition in AI voice assistants

The update arrives as competition in AI-powered voice assistants continues to grow. Anthropic recently enhanced Claude’s voice mode, enabling it to use Opus, Sonnet, and Haiku models to complete tasks across applications including Gmail, Calendar, Slack, Notion, and Canva, highlighting the industry’s push toward more capable AI assistants that can handle everyday workflows through natural conversation.