Voices

Where a voice comes from: a built-in preset, a voice designed from a description, or a cloned speaker.

Overview

A voice is the speaker POST /v1/audio/speech synthesizes with. Choose one in three ways: pass a preset id, create a voice from a written description, or clone a specific speaker from a reference clip. All three are spoken through chorus.

Send	Voice source	Guide
`voice` with a preset id	A built-in speaker	Voice presets
`voice` with the id of a designed voice	A voice created from a written description	Voice design
`ref_audio` and `ref_text`, or `voice` with the id of a cloned voice	A specific real speaker	Voice cloning

Design happens before the speech request. POST /v1/audio/voices/design returns previews, and POST /v1/audio/voices creates a voice from one. Cloning works either way: inline on each request, or as a voice created once on POST /v1/audio/voices and spoken by id.

Choose a voice source

Start with a preset. The speaker is fixed rather than generated, so it is the most consistent of the three.

Use design when no preset fits the character you want. A designed voice belongs to no real person, so there is no consent question to answer. Once created, it is the same speaker on every request.

Use cloning only when the voice has to be a particular person's. You need that person's audio and their permission. Send the clip inline with each request, or create a voice from it once and pass the voice's id.

List voices

GET /v1/audio/voices takes a model query parameter and lists its presets, followed by the voices your organization created.

curl "https://api.sprag.ai/v1/audio/voices?model=chorus" \
  -H "Authorization: Bearer $SPRAG_API_KEY"

The response is the OpenAI voice-list envelope. Each entry carries the id to pass as voice, a display name, a description, and a category: preset on a built-in speaker, designed on a voice created from a preview, cloned on a voice created from a clip. List the voices rather than hardcoding ids. The set can change over time.