Key facts
- OpenAI's ChatGPT desktop app now supports voice control for computer tasks.
- The feature uses OpenAI's new ChatGPT-Live voice models.
- Users can issue multi-step commands to control AI agents and applications.
- The update allows ChatGPT to access screen content on macOS via Appshots.
- Anthropic has also updated its Claude voice mode for app integration.
OpenAI announced Thursday that its ChatGPT desktop application has been updated to include ChatGPT Voice. This new functionality allows users to interact with the app using their voice to control AI agents and execute tasks directly on their computers.
The feature is powered by OpenAI's recently launched ChatGPT-Live family of voice models. The desktop version offers enhanced capabilities compared to its smartphone counterpart, enabling users to dictate complex, multi-step commands and respond to ChatGPT's prompts. A demonstration video showcased a developer using a single voice command to initiate a new thread, create a pull request, and identify the root cause of a software bug.
ChatGPT Voice integrates with both ChatGPT Work and Codex, and can also interact with websites and applications. On macOS, the Appshots feature allows the app to access the user's screen content, including alt-text. Users can also utilize ChatGPT Voice with Codex through the iOS app via remote access.
This development comes as competitor Anthropic has also enhanced its voice mode for Claude, allowing it to perform tasks within applications such as Gmail, Calendar, Slack, Notion, and Canva, leveraging its Opus, Sonnet, and Haiku models.