Speaker Identification

Speaker Identification labels each transcript segment with the person who said it. It runs entirely on your Mac as a post-recording step, applies only to audio captured while it's on, and it's off by default.

Turning It On

1. Open Settings → General.
2. Turn on Speaker Identification.

Because the feature creates a voiceprint for each speaker, enabling it asks you to confirm a short consent notice first. Voiceprints are biometric data, stored only on your device and never sent anywhere.

Once it's on, each recording is analyzed once you finish it - and imported audio when you bring it in - and its speakers are labeled in the transcript. It applies only to audio captured while the feature is on: existing recordings, and any parts recorded while it was off, stay unlabeled.

The Speakers Panel

Open a recording and click Speakers (the people icon) in the top-right toolbar, next to AI Chat. Each detected speaker is listed with:

• An initials avatar and name - unnamed voices start as Speaker 1, Speaker 2, and so on.
• A segment count showing how many transcript segments belong to that speaker.
• Row controls: the up arrow (Previous segment) and down arrow (Next segment) step through that speaker's lines in the transcript, and the pencil (Name speaker) names them.

In the transcript itself, every segment is prefixed with its speaker's avatar and name. If some parts of a recording have no labels, a Missing speaker labels? Learn why note at the bottom of the panel explains that audio captured while the feature was off isn't labeled.

Naming Speakers

• Click the pencil icon, or the speaker's name, to open the name field.
• Type a name and press Return to create a profile, or pick an existing person from the list to identify the speaker as them.
• The name is applied across all of that speaker's segments in the transcript.
• To undo, open the name field again and choose Untag - the speaker reverts to Speaker N.

Recognizing Familiar Voices

Once you've named someone, their voice profile is remembered and matched in your future recordings. When Silkwave Voice recognizes a familiar voice, the speaker row asks "Is this [name]?" - click Yes to apply the name in a single click, or Dismiss if it's wrong.

Managing Speaker Profiles

The people you name are saved as profiles you can manage in Settings → Speaker Profiles. The list shows everyone you've named across your recordings.

To remove a profile, click Delete. This unlinks it from every recording it was attached to, and those speakers revert to unnamed - it can't be undone. To remove the underlying voiceprints entirely, delete the recordings they were collected from.