Management · Saved Voices

Saved Voices API

Pre-approved organizations can register Korean and English voice samples and reuse them in DIVE saved mode.

Register voice

Send audio and an exact transcript for at least one language as multipart/form-data. Organizations without saved-voice approval receive E4101/403.

URL

POST

https://agitvxptajouhvoatxio.supabase.co/functions/v1/register-dive-voice-v1

Authentication and headers

Store API keys in server environment variables and send them only in request headers.

http

X-API-Key: {YOUR_API_KEY}
Content-Type: multipart/form-data; boundary=...

Request fields

namestringRequired
Name to save for the voice.
Up to 50 characters
descriptionstringOptional
Voice description.
Up to 200 characters
audioFileKoFileConditional
Korean audio file.
Required when audioFileEn is absent
audioTextKostringConditional
Exact transcript of the Korean audio file.
Paired with audioFileKo; must contain Korean
audioFileEnFileConditional
English audio file.
Required when audioFileKo is absent
audioTextEnstringConditional
Exact transcript of the English audio file.
Paired with audioFileEn; must not contain Korean

Register voice Request fields
Field	Type	Required	Description
`name`	string	Required	Name to save for the voice. Up to 50 characters
`description`	string	Optional	Voice description. Up to 200 characters
`audioFileKo`	File	Conditional	Korean audio file. Required when audioFileEn is absent
`audioTextKo`	string	Conditional	Exact transcript of the Korean audio file. Paired with audioFileKo; must contain Korean
`audioFileEn`	File	Conditional	English audio file. Required when audioFileKo is absent
`audioTextEn`	string	Conditional	Exact transcript of the English audio file. Paired with audioFileEn; must not contain Korean

Request example

bash

curl -X POST "https://agitvxptajouhvoatxio.supabase.co/functions/v1/register-dive-voice-v1" \
  -H "X-API-Key: {YOUR_API_KEY}" \
  -F "name=Customer support voice" \
  -F "audioFileKo=@voice-ko.wav" \
  -F "audioTextKo=Hello. This is the customer support voice."

Response

201

Registration result

idUUIDRequired
ID of the created voice.
namestringRequired
Saved name.
status"completed"Required
For a bilingual request, reference-token generation succeeded for at least one language. It does not mean both languages succeeded.

Register voice Registration result response fields
Field	Type	Required	Description
`id`	UUID	Required	ID of the created voice.
`name`	string	Required	Saved name.
`status`	"completed"	Required	For a bilingual request, reference-token generation succeeded for at least one language. It does not mean both languages succeeded.

Registration result

{
  "id": "550e8400-e29b-41d4-a716-446655440000",
  "name": "Customer support voice",
  "status": "completed"
}

400

E4201 / E4203 · Input or audio error

A multipart field is missing, a language transcript is invalid, or a decoded file is outside the 2–20 second range.

403

E4101 · Pre-approval required

The organization does not have access to saved voices.

503

E5002 · Voice file storage temporarily unavailable

The uploaded voice files could not be stored. The API does not return completed and removes the voice and files created by this request. If that cleanup also fails, the voice remains in failed status.

Accepted extensions are wav, mp3, m4a, webm, ogg, and flac; each audio file must be 2–20 seconds long.
The multipart request and the combined uploaded audio are each limited to 50 MiB.
The voice and its files are stored only after every uploaded file passes duration validation. Validation failures leave no voice or file behind.
For a bilingual request, status is currently completed even when token generation succeeds for only one language. The response has no per-language failure reason; the list exposes only englishAvailable, with no separate Korean-availability field.
A Korean-only registration attempts English generation in the background, so englishAvailable may be false immediately after completed. Background failures are not added to the registration response.
Legacy audioFile/audioText fields remain supported, but new integrations should use the per-language fields.

List voices

Lists the organization's saved voices in newest-first order.

URL

GET

https://agitvxptajouhvoatxio.supabase.co/functions/v1/list-dive-voices-v1

Authentication and headers

Store API keys in server environment variables and send them only in request headers.

http

X-API-Key: {YOUR_API_KEY}

Response

200

List response

voicesobject[]Required
List of id, name, description, englishAvailable, usageCount, status, and createdAt values.

List voices List response response fields
Field	Type	Required	Description
`voices`	object[]	Required	List of id, name, description, englishAvailable, usageCount, status, and createdAt values.

List response

{
  "voices": [
    {
      "id": "550e8400-e29b-41d4-a716-446655440000",
      "name": "Customer support voice",
      "description": null,
      "englishAvailable": false,
      "usageCount": 0,
      "status": "completed",
      "createdAt": "2026-07-28T00:00:00.000Z"
    }
  ]
}

Delete voice

Deletes an organization-owned voice and its stored reference files.

URL

DELETE

https://agitvxptajouhvoatxio.supabase.co/functions/v1/delete-dive-voice-v1/{id}

Authentication and headers

Store API keys in server environment variables and send them only in request headers.

http

X-API-Key: {YOUR_API_KEY}

Path parameters

idUUIDRequired
Saved voice ID to delete.

Delete voice Path parameters
Field	Type	Required	Description
`id`	UUID	Required	Saved voice ID to delete.

Response

200

Deleted

The response body is empty.

500

E5001 · Voice deletion failed

The stored voice files or the voice could not be deleted. If file deletion fails, the voice is kept.

The voice is deleted only after all of its stored voice files are removed. If file deletion fails, the API returns E5001 and keeps the voice.