Voice Cloning API

The Voice Cloning API accepts reference audio and text to synthesize in one multipart request, then returns audio. The reference transcript is optional.

API overview

Item	Value
Base URL	https://api.kitschlabs.com
Authentication	xi-api-key: <API_KEY>
Method	POST
Endpoint	/v1/voice-cloning/text-to-speech
Content-Type	multipart/form-data
Default output	mp3_44100_128 (audio/mpeg)
Additional output	wav_24000 (audio/wav)

How do I authenticate?

Send your issued API key in the xi-api-key header. Store the real key in a server environment variable or secret management service, and do not include it in browsers, app bundles, or logs.

xi-api-key: <API_KEY>

How do I generate speech?

When provided, reference_text must match the reference audio exactly. If it is omitted or blank, the service automatically transcribes the reference audio. The examples below send the reference file and text to synthesize in one request.

cURL

curl --request POST \
  "https://api.kitschlabs.com/v1/voice-cloning/text-to-speech?output_format=wav_24000" \
  --header "xi-api-key: <API_KEY>" \
  --form "text=This is the text to synthesize." \
  --form "reference_text=This is the exact reference transcript." \
  --form "reference_audio=@reference.wav" \
  --form "language_code=en" \
  --output speech.wav

Python

import os

import requests

with open("reference.wav", "rb") as reference_audio:
    response = requests.post(
        "https://api.kitschlabs.com/v1/voice-cloning/text-to-speech",
        params={"output_format": "wav_24000"},
        headers={"xi-api-key": os.environ["KITSCH_API_KEY"]},
        data={
            "text": "This is the text to synthesize.",
            "reference_text": "This is the exact reference transcript.",
            "language_code": "en",
        },
        files={"reference_audio": reference_audio},
        timeout=120,
    )

response.raise_for_status()

with open("speech.wav", "wb") as audio_file:
    audio_file.write(response.content)

JavaScript

import { readFile, writeFile } from "node:fs/promises";

const form = new FormData();
form.append("text", "This is the text to synthesize.");
form.append("reference_text", "This is the exact reference transcript.");
form.append("reference_audio", new Blob([await readFile("reference.wav")]), "reference.wav");
form.append("language_code", "en");

const response = await fetch(
  "https://api.kitschlabs.com/v1/voice-cloning/text-to-speech?output_format=wav_24000",
  {
    method: "POST",
    headers: {"xi-api-key": process.env.KITSCH_API_KEY},
    body: form,
  },
);

if (!response.ok) {
  throw new Error(`${response.status}: ${await response.text()}`);
}

await writeFile("speech.wav", Buffer.from(await response.arrayBuffer()));

Which fields do I send?

Field	Location	Type	Required	Description
text	form	string	Yes	Text to synthesize
reference_text	form	string	No	Exact transcript of the reference audio. Automatically transcribed when omitted
reference_audio	form	file	Yes	WAV, MP3, FLAC, or OGG file
language_code	form	string	No	One of en, ko, ja, or zh
instruct	form	string	No	Natural-language instructions for the speaking style
output_format	query	string	No	mp3_44100_128 or wav_24000

speed and duration cannot be set per request. The service applies the selected voice's default.

Reference audio can be up to 25 MiB and 20 seconds. When providing reference_text, include only the words that can be heard. Do not include timestamps, speaker names, subtitle markers, or translations.

How should I prepare reference audio?

• Use a clean, speech-only clip with one speaker.
• Avoid music, reverberation, background conversation, and strong noise.
• Do not cut off speech at the beginning or end of the clip.
• Choose a section that clearly contains the pronunciation and tone you need.
• Use only your own voice or a voice you have explicit permission to use.

Use only audio you have the right to use

Do not clone another person's voice without consent or use a cloned voice for impersonation, deception, or rights infringement.

Responses and errors

A successful request returns audio binary in the selected output_format. Save the response to a file and log the x-kitsch-request-id.

Status	Meaning	Recommended action
400·422	Required value, transcript, or file format error	Correct the request and reference audio
401	API key is missing or invalid	Check the key and xi-api-key header
402	Not enough available credits	Check the Billing status
413	Reference audio is too large	Reduce it to 25 MiB or less and retry
429	Account request limit exceeded	Follow Retry-After and apply backoff
5xx	Temporary server error	Record the request ID and retry a limited number of times