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

# Speakers in a transcript: rename, merge, filter

> How speech to text tells voices apart, and how to rename, add, merge and delete speakers, move a segment to another speaker and show one speaker's segments.

Speech to text tells the voices in a recording apart and gives each its own speaker. In the transcript editor you can give speakers real names, fix segments given to the wrong voice, merge two speakers that are the same person, and show one speaker's segments at a time.

## How speakers are found

When a transcript is ready, each voice the model tells apart becomes a speaker named Speaker 1, Speaker 2 and so on, in the order the model numbered them. If it can't tell voices apart, the transcript has one speaker, "Speaker".

Each speaker has a coloured avatar, and their name above each of their segments takes the same colour, so you can follow who is talking. The colour belongs to the speaker and stays when you rename them. There are six colours, so two speakers can share one.

The **Speakers** panel beside the transcript lists every speaker with their segment count. On a narrow window, open it with the **Speakers** button in the header.

![The Speakers panel with three speakers and their segment counts, Speaker 3 being renamed to Rana Aziz](https://mintcdn.com/hamsa-452fa79a/fdE-9_eEW1W6yh_s/images/media/speech-to-text/speakers-panel-light.webp?fit=max\&auto=format\&n=fdE-9_eEW1W6yh_s\&q=85\&s=0b1ab2fa870198472821251cb4120f9e)

## Rename a speaker

**1. Click the name**

In **Speakers**, click the speaker's name (**Rename speaker**). The field starts from the name shown, for example "Speaker 1".

**2. Type the new name**

For example "Dr. Layla Haddad" or "المذيع".

**3. Save**

Press Enter or click elsewhere. "Speaker renamed." confirms it, and every segment of that speaker shows the new name. Escape cancels.

A speaker name:

* can't be empty, and no other speaker in the transcript can already have it ("A speaker with this name already exists."),
* can't start with a number ("Start the name with a letter."),
* can't contain `@ # $ % ^ & * ( ) ! ? { }` ("Use letters, numbers and spaces only."),
* can be up to 250 characters.

## Add a speaker

Use this when the model missed a voice, so you can give segments to it.

**1. Open the form**

In **Speakers**, click **Add speaker**.

**2. Name it**

Type a **Name** (the same rules as a rename) and click **Add**.

![The Add speaker dialog with the Name Rana Aziz and Add](https://mintcdn.com/hamsa-452fa79a/fdE-9_eEW1W6yh_s/images/media/speech-to-text/add-speaker-light.webp?fit=max\&auto=format\&n=fdE-9_eEW1W6yh_s\&q=85\&s=ee570c45dab2dcd8f7215f2973dcda6e)

**3. Give it segments**

Edit each segment that belongs to the new speaker and pick them in the speaker list at the top of the form. See [Edit a segment](https://docs.tryhamsa.com/media/speech-to-text/transcript-editor.md#edit-a-segment).

## Move a segment to another speaker

Edit the segment and pick the right speaker in the speaker list at the top of the form. The change saves as soon as you pick, and only that segment moves.

## Merge two speakers

When the model split one person into two speakers, merge them.

**1. Pick the speaker to merge from**

Click **Merge speaker** on the speaker whose segments should move. It must have at least one segment.

**2. Pick the speaker to merge into**

"Now pick the speaker to merge into." Click **Merge speaker** on the other one.

![The Speakers panel after Merge speaker on Speaker 3, asking: Now pick the speaker to merge into](https://mintcdn.com/hamsa-452fa79a/fdE-9_eEW1W6yh_s/images/media/speech-to-text/merge-pick-target-light.webp?fit=max\&auto=format\&n=fdE-9_eEW1W6yh_s\&q=85\&s=6a41a06eff701b9d411717d1a22935ae)

**3. Confirm**

**Merge speakers?** says that every segment moves to the second speaker, and that the first one stays in the list with no segments. Click **Merge**. "Speakers merged." confirms it.

![The Merge speakers? dialog: every segment from Speaker 3 moves to Layla Hamdan, with Merge](https://mintcdn.com/hamsa-452fa79a/fdE-9_eEW1W6yh_s/images/media/speech-to-text/merge-speakers-dialog-light.webp?fit=max\&auto=format\&n=fdE-9_eEW1W6yh_s\&q=85\&s=f067f711105b93e2be5aa0a0070a897c)

Delete the empty speaker if you don't need it. See [Delete a speaker](#delete-a-speaker).

## Delete a speaker

Click **Delete speaker**, then **Delete** in the dialog (for example **Delete Speaker 3?**). Only a speaker with no segments can be deleted; otherwise you see "Move this speaker’s segments to another speaker first." Move or merge their segments, then delete.

## Show one speaker

In the toolbar above the transcript, the speaker button (**All speakers** until you pick one) opens **Show one speaker**: **All speakers**, then each speaker with their number of segments. Pick one to see only their segments; the button then shows their name. Click the X beside it (**Show all speakers**), or pick **All speakers**, to see everything again. The menu only appears when the transcript has more than one speaker.

![The Show one speaker menu under All speakers: All speakers, then Layla Hamdan, Karim Nassar and Speaker 3 with their segment counts](https://mintcdn.com/hamsa-452fa79a/fdE-9_eEW1W6yh_s/images/media/speech-to-text/show-one-speaker-light.webp?fit=max\&auto=format\&n=fdE-9_eEW1W6yh_s\&q=85\&s=02ceb70374b997506c196dfb57d8fc42)

## Speakers in exports and copies

**Copy with timestamp**, **Copy transcript with timestamps** and the SRT, WebVTT, Word and text exports use the names you see. PDF, web page and JSON exports use the names saved on the transcript: a speaker you haven't renamed appears there as `speaker_0`, `speaker_1`, and so on. See [Export a transcript](https://docs.tryhamsa.com/media/speech-to-text/export.md).

## Related

> [card] [Transcript editor](https://docs.tryhamsa.com/media/speech-to-text/transcript-editor.md)
>
> Edit segments and work from the keyboard.

> [card] [Export a transcript](https://docs.tryhamsa.com/media/speech-to-text/export.md)
>
> With or without speaker names.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.