ChatGPT Adds "Real-Time Dictation": Speak Into the Microphone and Speech Becomes Text Instantly
IT Home, November 19 — ChatGPT today announced a new "real-time dictation" feature: when users speak into the microphone, their speech is converted to text synchronously, with no need to wait until recording ends before transcription, allowing ChatGPT to receive voice instructions faster.

The feature will roll out gradually to all ChatGPT users worldwide, covering the web version, desktop client, and iOS and Android apps. To use it, users simply tap the audio waveform to view the real-time transcribed text, and after recording ends they can edit it directly in the input box.
OpenAI also announced that it is removing the original 10-minute duration limit on ChatGPT voice dictation and further improving transcription reliability. Earlier this year, ChatGPT's two-way voice conversation mode received a comprehensive overhaul.
Last year, OpenAI also launched a "recording mode" for transcribing voice notes and meeting content, though it remains in testing and is only available to Plus, Enterprise, Edu, Business, and Pro workspaces.
According to IT Home, users can also use the voice dictation feature in the ChatGPT input box to have ChatGPT convert voice messages into text, or organize them into bulleted key points.
Related Articles
- OpenAI Expands ChatGPT Free Academic Research Program: 50,000 More Spots Opening on November 23, Switching from Lottery to First-Come, First-Served
- Let Agents Help You Shop: OpenAI Launches Payment Feature ChatGPT Wallet
- Mac Version of ChatGPT Adds "Finance" Entry: Let AI Manage Your Books, Check Spending, and Track Investments
- OpenAI Announces Shutdown of ChatGPT's Old Voice Mode on December 11, Fully Moving to GPT-Live
- Study Shows OpenAI ChatGPT Accounts for Over 81% of Global AI Chatbot Traffic, Leading by a Wide Margin
- Reports of Sudden ChatGPT Account Bans, Even $200 Pro Users Affected