Editorial illustration for OpenAI adds voice to ChatGPT desktop, can now access apps and websites
ChatGPT Desktop Gets Voice Control and App Access
OpenAI adds voice to ChatGPT desktop, can now access apps and websites
OpenAI pushed an update to its ChatGPT desktop app on Thursday, bringing voice controls to a platform that previously only offered typing and clicking. The feature runs on ChatGPT-Live, the voice model family OpenAI launched earlier this month, and it works with both ChatGPT Work and Codex. Speak a command, and the app can now go fetch information from websites and apps rather than just answer from memory.
The desktop version outpaces what OpenAI shipped for smartphones. That earlier release handled interruptions well and made conversations feel more natural, but it stopped short of letting the app actually do anything on the phone. Desktop users get more: multi-step commands, mid-task check-ins when ChatGPT needs a decision, and on macOS, a feature called Appshots that lets the app see what's on screen, alt-text included.
OpenAI demoed the capability with a developer telling ChatGPT to spin up a new thread, file a pull request, and track down a bug's root cause, all from a single spoken instruction. Codex access extends to iOS through remote connection. OpenAI isn't alone in this push. Anthropic has been building out voice for Claude too, aimed at apps like Gmail, Slack, and Notion.
OpenAI on Thursday said it has updated its ChatGPT desktop app to add support for ChatGPT Voice, allowing users to talk to the app to control AI agents and perform tasks on their computer.
Why this matters For anyone building on top of ChatGPT, this is OpenAI moving voice from a novelty feature to a control layer for actual work. Talking to an agent that can pull up websites, read your screen through Appshots, and hand tasks to Codex changes what "voice interface" means on desktop, it's not a dictation tool anymore, it's an operator. Developers integrating ChatGPT Work or Codex into their workflows should expect users to start expecting voice control as a baseline, not an add-on.
Founders building competing agent products now have a bar to clear: OpenAI just bundled computer use, screen access, and voice into one desktop app, which narrows the gap between "chatbot" and "assistant that runs your machine." Researchers watching the ChatGPT-Live model family should pay attention to how OpenAI is pairing voice quality with tool use rather than treating them as separate tracks. The practical question worth watching is reliability: voice-driven agents that misread a screen or misfire a computer-use command carry more real-world risk than a chat reply gone wrong. How OpenAI handles error correction here will say more about readiness than the launch itself.
Related Reading
Browse all OpenAI news →
LLMs & Generative AI ChatGPT's 'Nerdy' tweak rewards goblin metaphors in answers, study finds
LLMs & Generative AI Google tests visual 'magazine-style' UI for Gemini 3 Pro users
LLMs & Generative AI AI Engineers Face Rising Costs, Need New Strategies for Efficiency
Related Topic OpenAI hires 630 ex-Meta staff as ChatGPT memory may turn data into ads
Business & Startups
Related Topic Sen. Warren urges Trump admin to expand AI credit after OpenAI letter
Business & Startups
From Archives Anthropic expands voice mode to Gmail, Slack apps
From Archives PhantomFill: When Language Models Invent Answers to Unanswerable Questions
From Archives ChatGPT Health Expands to All US Users, Adds Medical Record Integration
From Archives Single Tampered ChatGPT Link Spawns Rogue AI Agent in Minutes
Common Questions Answered
What new voice capabilities did OpenAI add to the ChatGPT desktop app?
OpenAI added voice controls to the ChatGPT desktop app powered by ChatGPT-Live, the voice model family launched earlier this month. Users can now speak commands to control AI agents and perform tasks directly on their computer, allowing the app to fetch information from websites and apps rather than just answering from memory.
How does the desktop voice feature differ from the smartphone version?
The ChatGPT desktop version with voice controls outpaces the earlier smartphone release, offering more advanced functionality and integration capabilities. The desktop version enables users to control AI agents and access apps and websites through voice commands, providing a more comprehensive experience than the mobile implementation.
Which ChatGPT products support the new voice control feature?
The new voice control feature works with both ChatGPT Work and Codex, allowing users to leverage voice commands across these platforms. This integration enables developers and users to incorporate voice as a control layer for actual work tasks rather than as a novelty feature.
How does voice control change the functionality of ChatGPT on desktop?
Voice control transforms ChatGPT from a dictation tool into an operator that can pull up websites, read screens through Appshots, and hand tasks to Codex. This shift means voice is no longer just for input but serves as a complete control layer for performing complex tasks and workflows on the desktop.
Why should developers expect voice control to become a baseline feature?
As OpenAI moves voice from a novelty feature to a functional control layer for actual work, users integrating ChatGPT Work or Codex into their workflows will increasingly expect voice control as standard functionality. Developers should anticipate that voice control will transition from an optional add-on to a baseline expectation in their applications.
Further Reading
- OpenAI brings its updated voice mode to the ChatGPT desktop app - The Verge
- OpenAI updating ChatGPT desktop app with GPT Voice for talking through work - 9to5Mac
- OpenAI adds ChatGPT Voice to desktop app, enabling computer control via voice - RuntimeWire
- OpenAI launch ChatGPT Advanced Voice mode on desktop and now PCs and Macs can join the conversation - TechRadar
- ChatGPT Brings Voice Mode Into Conversations, No Separate Interface - PCMag
