OpenAI brings ChatGPT Voice to desktop app
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- OpenAI updated its ChatGPT desktop app Thursday to add ChatGPT Voice, powered by the ChatGPT-Live voice model family launched earlier this month.
- The desktop version lets users dictate complex multi-step commands and respond when ChatGPT needs input — a capability the original smartphone launch lacked, where voice only handled smoother conversation and better interruption, not action.
- ChatGPT Voice integrates with ChatGPT Work and Codex, taps computer-use skills to browse websites and apps, and on macOS uses Appshots to read screen content including alt-text.
- In a demo video, OpenAI showed a developer issuing one voice command that created a new thread, opened a pull request, and identified the root cause of a bug.
- Codex users can trigger ChatGPT Voice remotely through the iOS app.
- Anthropic simultaneously updated its Claude voice mode to tap Opus, Sonnet, and Haiku models across Gmail, Calendar, Slack, Notion, and Canva.
Why it matters: OpenAI is repositioning voice from a conversation layer into an agent controller on the desktop — the same single-command demo that opened a pull request is the kind of workflow Anthropic is targeting across productivity apps, so the competitive gap is now defined by how many ecosystems each voice agent can actually act inside.
