Voice Selection

Finding a Speaker IDCopied!

To use the Replica API for most speech generation tasks, you'll need a speaker ID.

Although it might not seem obvious at first, voices on Replica are grouped by several different IDs with the speaker ID representing the combination of both the desired voice and style.

The required information is available on the /library/voices endpoint.

Example API request to get available voices:

import requests

url = "https://api.replicastudios.com/v2/library/voices"

headers = {
"X-Api-Key": "..."  # Replace with your API key
}

response = requests.get(url, headers=headers)
voices = response.json()

The response will include voice details including:

UUID (Do note this is the voice's UUID, not quite what we're after)
Name
Available styles
Voice characteristics
Sample audio URLs

For example, here is a truncated example of "Freya" from the library:

{
    "uuid": "2bfc6875-308c-4101-bf4d-7c279bc56db2",
    "name": "Freya",
    "avatar_url": "https://static.replicastudios.com/production/uploads/avatars/2bfc6875-308c-4101-bf4d-7c279bc56db2.jpg",
    "hero_image_url": "https://static.replicastudios.com/production/uploads/avatars/2bfc6875-308c-4101-bf4d-7c279bc56db2_heroimage.png",
    "hero_sample_url": "https://static.replicastudios.com/production/uploads/previews/2bfc6875-308c-4101-bf4d-7c279bc56db2_herosample_1698884342198.wav",
    "description": "The female action lead. Whatever your storyline in the realm of the magical or medieval, Freya will blast through or come swinging in to the rescue. With a wide voice range, she'll challenge any man, monster or machine. ",
    "metadata": {
        "accent": "British",
        "gender": "female",
        "voiceAge": "youngadult"
    },
    "characteristics": [
        "Authoritative",
        "Brave",
        "Clear",
        "Commanding"
    ],
    "default_sample": null,
    "default_style": {
        ...
    },
    "styles": [
        {
            "uuid": "07e62901-72c4-46e5-b009-aa0938d749df",
            "name": "Serious",
            "description": null,
            "speaker_id": "9b1f5c24-a18b-4b9e-a785-b3a3b3b8751a",
            "avatar_url": "https://static.replicastudios.com/production/uploads/avatars/07e62901-72c4-46e5-b009-aa0938d749df_style_avatar.jpg",
            "samples": [
                {
                    "uuid": "cee6c584-af5a-4b8f-84cf-06723d887c54",
                    "text": "You can’t keep running from this. Sooner or later you’ll have to face the truth.",
                    "url": "https://static.replicastudios.com/production/uploads/previews/5ba7b998-cba8-4ec8-9524-ca69e06ae4de.mp3"
                }
            ],
            "capabilities": {
                "tts.vox_1_0": true,
                "tts.vox_2_0": false,
                "sts.vox_1_0": true,
                "sts.vox_2_0": false
            }
        },
        ...
    ]
}

If you then select the style you want, you'll see the speaker_id as a field under the style.

In the above example, the speaker_id is 9b1f5c24-a18b-4b9e-a785-b3a3b3b8751a for the style "Serious".

Voices also have a default_style field, which (when not truncated) will have the same format as styles.

If you're unsure which style to use, you can use the default_style field to get a feel for the voice.

Finding a Performer Style IDCopied!

Each voice (or 'performer' in this case) can have multiple styles. When querying the /library/voices endpoint, you'll receive style information in the response including:

Style UUID
Style name
Speaker ID
Sample audio for that style
Capabilities (supported model versions)

The style UUID is used as the performer_style_id parameter when making speech-to-speech generation requests.

Using a Voicelab Voice PresetCopied!

What is Voicelab? Learn more about it here.

Voice presets allow you to save specific voice configurations for reuse. When making a speech generation request, you can include the voice_preset_id parameter alongside the speaker_id parameter.

Your voice presets are available in the /my/voices endpoint, alongside style/speaker IDs that will work alongside it.

Voice presets are only available on certain subscription tiers. Check your account settings for availability.

Previous Page

Authentication

Next Page

Global Controls