Skip to main content
Speaker attribution in Zoom is available in the Granola desktop app on macOS. It uses the Zoom Workplace desktop app and macOS Accessibility access; it does not require the Granola browser extension.

​
Overview

Speaker attribution helps Granola identify who is speaking in a Zoom call and show their Zoom display name in your transcript. Instead of seeing every other participant as Them, you may see each person’s name above their parts of the conversation. Your own speech continues to appear as Me. While Granola is transcribing a call in the Zoom desktop app, it uses the active-speaker information and participant names shown by Zoom to match speech to transcript sections.

​
Before you start

You will need:
  • The Granola desktop app on macOS
  • A call in the Zoom Workplace desktop app
  • macOS Accessibility access granted to Granola
Zoom speaker attribution is not currently available on Windows or in the Zoom web client. It does not require the Granola browser extension. Speaker attribution only applies while Granola is actively transcribing a Zoom call. It does not add names to older transcripts.

​
Set up speaker attribution

Granola will guide you through granting the macOS permission needed for Zoom speaker attribution.
  1. Open the Granola desktop app.
  2. If needed, manually refresh Granola by pressing Cmd+R.
  3. Go to Settings > Preferences and click Speaker attribution, at the bottom of the General section.
  4. Go to Settings > Preferences > Speaker attribution.
  5. Next to Zoom, click Set up.
  6. Turn on Accessibility access for Granola, then return to the Granola app.
The Zoom row will show Connected once Granola can access the information it needs from Zoom. There is no separate Zoom speaker-attribution toggle.

​
About Accessibility access

Accessibility is a macOS system permission that allows apps to read and interact with interface elements in other apps. For Zoom speaker attribution, Granola uses it to read the participant display names and active-speaker indicator shown in the Zoom desktop app. Speaker attribution does not use Accessibility access to record your screen, capture audio, or access your camera or microphone. Granola captures audio for transcription separately, as described in How transcription works.

​
Use speaker attribution in a meeting

Start your call in the Zoom Workplace desktop app and start transcription in Granola as normal. When Granola can identify another person speaking through Zoom, their Zoom display name appears above their transcript bubbles. Your own speech remains labelled Me. For what speaker attribution can and cannot identify, and how it relates to privacy and in-meeting notice, see the Speaker attribution overview.

​
Troubleshooting

This feature is currently only available on macOS. First go to **Settings > Preferences > Speaker Attribution.If it is not visible , Zoom speaker attribution has not yet been enabled for your account or workspace.
  1. Open System Settings > Privacy & Security > Accessibility.
  2. Make sure Granola is listed and switched on.
  3. Return to Granola and reopen Settings > Preferences > Speaker attribution.
  4. If the Zoom row still does not show Connected, fully quit and reopen Granola.
If Granola is already switched on in Accessibility settings but the permission is not detected, switch it off and on again, then reopen Granola.
Check that:
  • You are using the Zoom Workplace desktop app on macOS, rather than the Zoom web client.
  • The Zoom row under Settings > Preferences > Speaker attribution shows Connected.
  • Granola was open and transcribing while the Zoom call was active.
  • The transcript was captured after speaker attribution was set up; names are not applied retrospectively.
Some transcript sections may not have a speaker name if Zoom did not show a clear active speaker at that moment, participants were speaking over each other, or several people were using a shared meeting-room device.
Speaker attribution relies on the participant name and current-speaker information shown by Zoom. If a participant uses a different display name, several people use one room device, or people speak over each other, the label may not identify the individual correctly.