> ## Documentation Index
> Fetch the complete documentation index at: https://docs.bowerlabs.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Speaker labels for meeting recordings

> Record a meeting and Bower labels each speaker's part of the transcript.

**Shipped 2026-06-29 · Voice notes**

When you record a voice note with more than one person talking, Bower can now
identify the different speakers and label each turn of the conversation — so a
meeting transcript reads as a clear back-and-forth instead of one undivided block
of text.

## What's new

* A **Meeting** toggle next to the voice recorder. Turn it on before recording a
  conversation with multiple speakers.
* The transcript is split into labelled turns — **Speaker 1**, **Speaker 2**, and
  so on — each as its own paragraph.
* Solo dictation is unchanged: with the toggle off (the default), your note is
  transcribed as a single flowing transcript, exactly as before.

## How it works

With **Meeting** on, Bower asks the speech engine to separate the voices in the
recording and tags every passage with the speaker who said it. Your transcript is
then formatted with each speaker's turn on its own line. The speaker labels are
attached after transcription and never rewritten during clean-up, so attribution
stays accurate.

## Good to know

* Speakers are labelled generically (**Speaker 1**, **Speaker 2**, …). You can
  rename them in the note.
* Works best for clear recordings; very large groups (4+ speakers) or noisy audio
  may reduce accuracy.
* Available for recorded and uploaded voice notes. Live mode is not yet supported.
* Supported for English and 14 other languages.