If your voice-to-text button does nothing, or if the browser shows a blocked mic icon in the address bar, Chrome is silently denying microphone access. This guide walks through both the site-level fix and the system-level fix so dictation starts working in about a minute.
Fix the site-level permission
Most voice tools ask for microphone access the first time you press record. If you dismissed that prompt or clicked block, Chrome remembers the choice and will not ask again until you clear it.
Open the site (for example zahvox.com/tool) in Chrome.
Click the small tune or lock icon on the left of the address bar.
Find Microphone and switch it to Allow.
Reload the page and press the record button again.
Fix the browser-wide setting
Chrome also has a global microphone control. If the whole browser is muted, no site can use dictation regardless of its own permission.
Open chrome://settings/content/microphone in a new tab.
Set the default to sites can ask to use your microphone.
Confirm the correct microphone is selected at the top.
Remove any site listed under Not allowed that should be allowed.
Fix the operating system permission
On macOS and Windows the operating system can block Chrome from touching the microphone at all. When that happens the browser prompt appears normal but no audio ever reaches the page.
macOS: System Settings, Privacy and Security, Microphone, enable Google Chrome.
Windows 11: Settings, Privacy and security, Microphone, allow apps and desktop apps, then confirm Chrome is on.
Restart Chrome fully after changing the OS setting so it can re-read the permission.
Verify it works
Once permissions are set correctly you should see a small red dot in the tab and a live recording indicator in the address bar as soon as you press record. If not, try a hard refresh with Ctrl or Cmd plus Shift plus R and pick a different microphone from the input dropdown if the site offers one.
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 chrome 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.
