Recording all your meetings has never been easier than with the Fellow Note Taker! Fellow's AI Note Taker identifies speakers in your meeting recordings and attributes their words to them in the transcript, whether everyone is joining remotely or some attendees are sharing a room. This article covers how speaker identification works across different meeting types, how to edit speaker labels in the transcript, and how to manage voice detection settings.

Each speaker is automatically identified and attributed to the microphone they are using. No additional setup is required.

When multiple people share a microphone in the same room, Fellow detects that multiple voices are present and attempts to identify each speaker individually.

If speakers cannot be identified, generic labels will be assigned, such as:

Fellow can automatically assign names based on past meetings, but only for users in your workspace with voice recognition enabled. Please note that this feature does not apply to external attendees.

If a speaker is not automatically identified or is mislabeled (pictured below), you can update their name directly in the transcript.

To update these to the correct speaker, you'll need to:

Voice detection is what allows Fellow to recognize and remember speaker voices across meetings. It can be managed at both the workspace level (by admins) and the individual user level.

Workspace admins can enable or disable voice detection for all users via Workspace Settings > Note Taker & AI > AI:

Individual users can enable or disable voice detection for their own account by going to User Settings > Note Taker > Voice matching:

Once enabled, it will look something like this:

You can then click into any audio sample to review or add speaker entries.