Skip to main content
GET
/
v1
/
voices
/
search
Search voices.
curl --request GET \
  --url https://supertoneapi.com/v1/voices/search \
  --header 'x-sup-api-key: <api-key>'
{
  "items": [
    {
      "voice_id": "<voice-id>",
      "name": "Agatha",
      "age": "young-adult",
      "gender": "female",
      "use_case": "narration",
      "use_cases": [
        "narration",
        "storytelling"
      ],
      "language": [
        "ar",
        "bg",
        "cs",
        "da",
        "de",
        "el",
        "en",
        "es",
        "et",
        "fi",
        "fr",
        "hi",
        "hu",
        "id",
        "it",
        "ja",
        "ko",
        "nl",
        "pl",
        "pt",
        "ro",
        "ru",
        "vi"
      ],
      "styles": [
        "kind-default",
        "normal",
        "serene"
      ],
      "models": [
        "sona_speech_1",
        "sona_speech_2",
        "sona_speech_2_flash",
        "supertonic_api_1",
        "supertonic_api_3"
      ],
      "description": "",
      "samples": [
        {
          "language": "ko",
          "style": "kind-default",
          "model": "supertonic_api_3",
          "url": "https://example.com/samples/sample-audio.wav"
        }
      ],
      "thumbnail_image_url": "https://example.com/thumbnails/voice-thumbnail.png"
    }
  ],
  "total": 150,
  "next_page_token": "some_opaque_token_string_representing_last_id"
}
Returns preset voices matching the given filters. Multiple values per parameter are comma-separated and treated as OR conditions.

​
Endpoint

GET https://supertoneapi.com/v1/voices/search

​
Query parameters

NameDescriptionExample
nameVoice name, partial match.Coco
descriptionDescription, partial match.kind and gentle
languageLanguage code, comma-separated.ko,en,ja
genderVoice gender, comma-separated.male,female
ageVoice age group.child, young-adult, middle-aged, elder
use_casePrimary use case.audiobook, narration, advertisement
use_casesRecommended use cases (OR).audiobook,narration
styleEmotional style. The first value is the default.neutral, happy, sad, angry
modelSupported model.sona_speech_1
page_sizeItems per page. Default 20, max 100.50
next_page_tokenToken from a previous response.eyJpZCI6IjEyMzQ1In0=
sort is not supported. Results return in their natural order; sort client-side if needed.

​
Examples

Filter by style and language: 
GET /v1/voices/search?style=happy&language=ko,en
Returns voices that include happy in their style list and support either Korean or English. Paginate: 
GET /v1/voices/search?page_size=50&next_page_token=eyJpZCI6IjEyMzQ1In0=

​
Response

Returns items (array of voice objects) and next_page_token (string, optional). If no voices match, items is an empty array — the request still returns 200 OK.

​
Notes

  • Comma-separated values are OR-ed: language=ko,en&style=happy,sad returns voices that support Korean OR English and have happy OR sad in their styles.
  • Empty results are not an error — check items.length rather than relying on a 4xx response.
  • For the full voice object shape (samples, supported models, etc.), see Voices.

​
See also

Docs: Voices

Concept walkthrough with SDK code.

Voice search example

Build a searchable in-app voice picker.

Authorizations

​
x-sup-api-key
string
header
required

Query Parameters

​
page_size
number

Number of items per page (default: 20, min: 10, max: 100)

​
next_page_token
string

Token for pagination (obtained from the previous page's response)

​
name
string

Search across name. Space separated.

​
description
string

Search across description. Space separated.

​
language
string

Filter by language (comma-separated)

​
gender
string

Filter by gender (comma-separated)

​
age
string

Filter by age (comma-separated)

​
use_case
string

Filter by use case (comma-separated)

​
use_cases
string

Filter by use cases array (comma-separated for OR logic)

​
style
string

Filter by style (comma-separated for OR, semicolon-separated for AND). Mixing comma and semicolon is invalid and will result in 400. Note: AND semantics apply across styles on a single character; cloned voices have a single style and will only match AND when exactly one style is requested and equals the cloned voice style.

​
model
string

Filter by model (comma-separated)

Response

Paginated available voices response with next page token

​
items
object[]
required

List of character items

​
total
number
required

Total number of available characters (might be approximate or removed in future)

Example:

150

​
next_page_token
string

Token for fetching the next page of results. Undefined if no more pages.

Example:

"some_opaque_token_string_representing_last_id"