ChatGPT adds real-time Live Dictation and kills the 10-minute recording limit

OpenAI just made ChatGPT dictation stream live as you speak. The 10-minute cap is gone, and it is rolling out to all plans globally.

Tools · Source: 9to5Mac

What happened

OpenAI just upgraded ChatGPT with a massive improvement to its speech-to-text capabilities. The new feature is called Live Dictation. Your speech now turns into text instantly as you talk. You no longer have to wait for the app to process a finished audio recording. The update rolled out globally on November 18.

The company also killed the old 10-minute recording limit. Atty Eleti from the OpenAI voice team confirmed the cap removal as part of this streaming upgrade. You can now speak for as long as you want without the app cutting you off. Users can tap or click the waveform on their screen to show or hide the text as it generates in real time.

This feature is not locked behind a premium paywall. Live Dictation is rolling out to all ChatGPT plans, including the free tiers. It works across iOS, Android, the web, and the desktop app. OpenAI President Greg Brockman called voice a superior interface for computers. He noted that live dictation adds significant ergonomic benefits for users who want hands-free control.

Key facts

Why it matters

This changes how founders and engineers interact with AI on the go. The old record-then-wait flow constantly broke user momentum. Now you can brain-dump product specs, draft long emails, or talk through complex bugs without hitting a hard stop at the 10-minute mark. Real-time visual feedback means you catch transcription errors immediately instead of after a massive monologue. You spend less time waiting and more time actually building. For busy professionals, this removes the friction of typing out long prompts on a mobile keyboard.

The second-order effect is a massive industry shift toward voice-first workflows. As dictation becomes frictionless and ergonomic, typing becomes the fallback rather than the default. Teams will start integrating conversational tools deeper into their daily operations. Developers building AI applications will need to match this new baseline for speed and reliability. If your app still makes users wait for a spinning loading wheel after they speak, your product will feel broken by comparison. The standard for voice interfaces just got much higher.

For builders

Voice interfaces are the new baseline

Users now expect real-time transcription without arbitrary time limits. If your AI product uses a record-then-wait flow, you will lose users to competitors who stream text instantly. You must upgrade your audio processing pipelines.

Productivity tools must adapt

Professionals will increasingly use voice to draft long documents and code logic. Builders who integrate seamless voice-to-text APIs into their workflow tools will capture this shifting user behavior. Those who rely solely on text input will fall behind.

Ergonomics drive user retention

Greg Brockman highlighted the ergonomic benefits of voice. Hands-free operation keeps users in your app while they commute or walk. Founders who optimize for mobile voice input will see higher engagement metrics.

My take

Typing is rapidly becoming a legacy input method. I build products every day, and hitting a 10-minute wall while brain-dumping an architecture idea was incredibly frustrating. OpenAI removing this friction proves that the future of human-computer interaction is just talking to your machine. Stop typing and start talking.

Original reporting: 9to5Mac. This is my rewrite and opinion.

More AI news for builders