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

# Voices and avatars

> The voices Goosy can speak your videos in, and the presenters he can put on camera.

Two registries decide who your videos sound like and who they look like.
**Voices** holds the voices Goosy can narrate in. **Avatars** holds the
on-camera presenters he can use. Both live in **Settings**, one under the other,
and both work the same way: a list of what this workspace already has, plus two
ways to add more.

<Warning>
  Both screens carry the same standing rule, and it is not a formality. Create a
  voice clone or use a likeness **only with that person's recorded consent**.
</Warning>

## Voices

**Settings → Voices** is described on the screen as *the voices your assistant
can speak in — cloned or licensed*. Under the heading, **Voice registry** shows
how many you have.

### Reading a row

Each row gives you the name, then the provider and what kind of voice it is —
whether it came ready-made, is a professional voice, or is one you added — and
any note that was saved with it.

On the right of the row is a status, and a voice with nothing to flag carries
no badge at all — it is simply usable.

| Status             | What it means                                                                                                 |
| ------------------ | ------------------------------------------------------------------------------------------------------------- |
| **Default**        | Usable, and set as the default for at least one kind of production.                                           |
| *(no badge)*       | Usable. Approved, and the rights behind it are cleared, owned or licensed.                                    |
| **Pending review** | Not usable yet, and still moving — either the voice is not approved, or the rights are still being checked.   |
| **Unavailable**    | Not usable, and it will not clear on its own — the voice has been switched off, or its rights were withdrawn. |

Only a usable voice can be previewed, or picked when something is being made
— the badge and the availability are the same answer.

### Adding a voice

Two controls sit in the header, and they do different jobs.

<Steps>
  <Step title="Pick from library">
    Opens **Add a voice from the library** — audition a curated default, or
    search the full library. **Defaults** is a shortlist; **Browse voices** has a
    search box and tags each result with its type, the speaker's rough age and
    accent, and what it suits — conversational, narration, social, and so on.
    Anything already in your registry is marked **Added**; everything else
    offers **Use this voice**.
  </Step>

  <Step title="Add a voice">
    For a voice you already hold with the provider. Give it a **label** you will
    recognise, paste the **voice ID** from your provider account, and optionally
    set the voice type, the language as a standard language tag such as
    `en-US`, and a short note on what it is for.

    The ID is taken exactly as you paste it — it is not checked with the
    provider first — so copy it rather than typing it from memory.
  </Step>
</Steps>

### Changing one later

**Manage** on any row offers **Edit** and **Deactivate**. If you do not have
permission to manage this workspace's settings, the control is not there at
all.

Editing changes the label, the type, the language and the note. It does not
change which voice it points at: the provider ID is fixed once a voice is added,
so to point somewhere else, **Deactivate** the old entry and add a new one —
there is no separate remove; a deactivated entry stays in the list.

Voice credentials are stored encrypted.

## Avatars

**Settings → Avatars** holds *your on-camera likeness for avatar-led videos*. It
reads the same way as Voices — a list, two ways to add, **Manage** on each row.

### Reading a row

Name, provider, and any note saved with it. The status is about consent and
rights, and it always names the actual reason rather than a general "not
ready":

| Status              | What it means                                                                   |
| ------------------- | ------------------------------------------------------------------------------- |
| **Ready**           | Usable now — active, consented, rights cleared, and the consent has not lapsed. |
| **Consent pending** | The person has not yet given consent.                                           |
| **Rights pending**  | Consent is in, the rights are still being checked.                              |
| **Consent expired** | Consent has lapsed and needs to be given again.                                 |
| **Consent revoked** | Consent was withdrawn. This does not clear on its own.                          |
| **Rights revoked**  | The rights were withdrawn. Also permanent.                                      |
| **Inactive**        | Switched off deliberately.                                                      |

A row whose consent is still valid but lapses within the next month is flagged
as expiring soon, so you can renew it before a render fails.

### Adding an avatar

<Steps>
  <Step title="Pick from library">
    **Add an avatar look** — preview a curated default, or browse the public
    library on the **Browse avatars** tab. If nothing is curated for you yet, the
    Defaults tab says so and points you at Browse.
  </Step>

  <Step title="Add an avatar">
    For an avatar you already hold with the provider. Give it a **display name**,
    paste the **avatar ID** from your provider account, optionally record how it
    was created, and choose where it can be used — as the presenter in an avatar
    lesson, as the picture-in-picture face over a screen share, or both.

    As with voices, the ID is pasted exactly and is not verified with the
    provider first.
  </Step>
</Steps>

**Manage** offers **Edit** and **Deactivate**, the same as Voices.

An avatar added by hand is available to your whole account — sibling workspaces
can pick it too, so treat a likeness as shared once it is in. The renders made
with it stay keyed to the workspace that made them.

## Where these get used

A production profile can name a default voice and a default avatar for the kind
of video it makes. Those defaults are provisioned values — you can see them on
**[Production profiles](/setup/production-profiles)**, but there is no control
there for changing them today.

## What to do next

* **[Strategy and taste](/setup/strategy-and-taste)** — steering what Goosy
  writes without repeating yourself.
* **[Production profiles](/setup/production-profiles)** — the look and shape a
  render comes out in.


## Related topics

- [Production profiles](/setup/production-profiles.md)
- [Brand](/setup/brand-profile.md)
- [Glossary](/help/glossary.md)