Voice cloning from ten seconds in the request.
0
training jobs between a clip and a voice
No enrollment pipeline, no per-voice fee, no waiting for a training job. The reference clip is a request parameter, and the identity it carries survives translation.
The tape, voice specimens, pre-rendered on the production API
generated 2026-08-05
Leader · The whole enrollment
Five to ten seconds of one consenting speaker, up to ~1.5 MB of WAV. The clip becomes a request field; references are fingerprinted and cached, so the first call pays the cloning work.
import { readFileSync } from "node:fs"
// ten seconds of one consenting speaker, the whole enrollment
const wav_b64 = readFileSync("reference.wav").toString("base64")
const body = {
transcript: "Same identity, any of 23 languages.",
language: "de", // a language the
voice: { mode: "clone", wav_b64 }, // reference never spoke
}
// fingerprinted + cached: the first request pays the cloning
// work, every reuse of the same clip skips it
The takes, every scenario, one deck
An audiobook chapter, a dubbed film line, an e-learning module, long-form narration where the read has to hold for hours, not seconds.
“The lighthouse keeper counted ships the way other people counted sheep, and by the end of most winters he could tell a trawler from a tanker by the sound of it alone, without ever leaving his chair. He had kept the light for nineteen years. He had spoken to perhaps forty people in that time, and remembered all of them. On the night this story begins, the sea was flat, the lamp was turning, and one of the ships was counting back.”
Run-out · Provenance, stated
- Every synthesized sample carries an inaudible watermark, applied at generation on every request.
- Customer reference audio never enters a training set.
- Usage is per-key and auditable at GET /v1/usage. The full policy sheet is at provenance and privacy.
Notes
What makes a good reference clip?
Five to ten seconds of one speaker talking naturally in a quiet room. Clean beats long, the cap is ~1.5 MB of WAV, about 30 seconds, and extra length past ten seconds buys little.
Do I pay per cloned voice?
No. There is no per-voice fee and no voice limit. You can carry one clone per end customer without a pricing conversation.
A key, one stream, your own script, nothing on it counted while you build.
Get a key, run your own script