Use speaker labels in a recording (Pixel 6 & later, including Fold)

Important: Speaker labels are available only in English.

In a recording transcript, you can use speaker labels to know when different people are speaking. The Recorder app will auto-detect different speakers and label them as Speaker 1, Speaker 2, Speaker 3, and so on.

Turn speaker labels on & off

On your device, open the Recorder app .
Tap the user icon .
Tap Recorder settings Speaker labels.
When prompted, tap Yes, enable.
To turn off Speaker labels, go back to Recorder settings.

Update speaker labels

From “Transcript view,” tap Edit .
Select the dropdown menu beside the speaker label you want to change.
To apply changes for the current paragraph, tap Rename, Create a new speaker, or other labels.
To make the change, tap Save.

Tip: If you change a speaker label for an individual paragraph in the transcript, the same speaker labels in other paragraphs won’t change.

Split a paragraph to update speaker labels

For any recording, tap Edit Split from here.
You can create 2 sections from a single paragraph and make further changes to speaker labels.

Search speaker labels

You can search for people in a transcript with the search function on the Recorder home screen:

Tap Search .
Type the name of the person you're looking for.
Scroll through the list of all search results.

You can also search speaker names in a specific recording:

Tap on an existing recording.
Near the bottom right, tap Transcript.
Tap Search .
Type the name of the person you're looking for.

Share speaker labels along with transcript

Speaker labels are visible when you share your recording transcript through a video clip, text file, or a Google Doc. Learn how to share your recordings.

Learn how Recorder handles your data for speaker labels

If you turn on speaker labels, Recorder analyzes your audio recordings to detect and label distinct speakers within your transcripts. To do this, Recorder:

Computes models of the voices detected in the recording on the device.
Gives each voice detected a generic text label such as Speaker 1, Speaker 2, and so on.
Compares the voices throughout the recording so the transcript can reflect what each voice said and when.

These steps happen on your phone or tablet and the generated speaker text labels are saved in your transcript.

Tip: The voice models mentioned above may be considered biometric data in some countries or regions.

Learn about voice models retention & deletion policy

The voice models used to generate speaker labels are stored temporarily on your device and deleted within a few minutes after labeling is complete.