Speakers are separated incorrectly
In shortThe AI separates speakers automatically and can be wrong. In the recording, tap a speaker label: you can name the person, merge two speakers into one, or change the speaker of a single line.
How to fix the speakers
iPhone and iPad
- Open the recording on the Text tab.
- Tap the speaker label above a line (for example, “Speaker B”).
- Choose an action: rename, Someone else said this or All speakers in this recording.
Android
- Open the recording on the Text tab.
- Tap the speaker label above a line (for example, “Speaker B”).
- Choose an action: rename, Someone else said this or All speakers in this recording.
Web
- Open the recording. The text is on the left; on a narrow screen it is the Text tab.
- Click the speaker’s name next to a phrase (for example, “Speaker B”). A card opens; Voice plays a sample of that speaker when the recording has audio.
- Type a name and click Rename everywhere. Same person as… merges this speaker into another one; Only this fragment gives one phrase to someone else or to + New speaker.
What each action does
- Rename replaces “Speaker B” with a name everywhere in this recording. Keep the name short: the app will not save one that is too long.
- Someone else said this changes the speaker for the chosen line only.
- This is the same person as… merges two speakers when the AI split one person in two. The app asks “Merge speakers?”.
After renaming, the app offers “Rebuild the notes with names?”. The web version makes the same offer on the Notes tab: Rebuild with names.
What does not change
The SRT subtitle file built by the server keeps letters instead of names. TXT and VTT files exported from the web version carry the names.
If the speaker plate is missing
The app says “Speaker names aren’t available yet”. Update SozAI, and if the message stays, Contact us.
Did this help?Thank you for the feedback.