Guide · SetupAvailable now

Enable microphone permissions in Safari

Safari has strong privacy defaults, which means voice-to-text sites are blocked until you explicitly opt in. Here is the quick fix on both Mac and iPhone.

Try the free voice toolFree · No signup · Runs in your browser

Safari does support the Web Speech API on recent versions, but the microphone permission is off by default and hidden inside Preferences. This guide covers both macOS Safari and iOS Safari so you can dictate from any Apple device.

Allow the site on macOS Safari

Open the voice-to-text page in Safari.
In the menu bar choose Safari, Settings, Websites.
Pick Microphone in the left sidebar.
Set the site to Allow.
Reload the page.

Allow the site on iPhone or iPad Safari

Open Settings on your device.
Scroll down and tap Safari.
Tap Microphone.
Set it to Ask or Allow.
Return to the site and press record. Safari will prompt once.

Check macOS system permissions

macOS keeps its own list of apps allowed to use the microphone. If Safari is not on it, the browser prompt appears but no audio arrives.

Open System Settings, Privacy and Security, Microphone.
Turn on Safari.
Close and reopen Safari for the change to take effect.

Know the limits of Safari speech

Safari support for continuous voice-to-text is more restricted than Chrome or Edge. Long sessions can time out, and some sites detect Safari and fall back to a manual mode. If you rely on dictation daily, keep Chrome or Edge as a backup and use Safari on iPhone for quick captures.

Build a daily voice-to-writing habit

The first week is about proving the setup works. The second week is about turning it into a habit. Pick two moments in the day where enable microphone permissions in safari would save the most time, and commit to using voice for those moments only. Once the muscle memory is there, expand to a third and fourth surface. The goal is a workflow you reach for without thinking, not a demo you show off once and forget.

Where Zahvox fits in this workflow

Zahvox is the careful pass that sits after any raw capture. Once you have the words on the page, paste them into the free tool at /tool, tighten the wording, and copy the result back. That extra minute is what separates a dictation dump from real writing, and it works the same way whether the source was a browser, a native app, or a phone keyboard.

Looking for something else? Browse the full guides library or open the free voice tool and practice as you read.