CosyVoice Clone

Generate lifelike speech from text using CosyVoice with optional reference prompts for instant cloning.

Backend online

Output Text

0 / 1000

Reference Script

We recommend a short sample (1–2 sentences) matching your uploaded reference.

Prompt Audio

🎙️

Drop a reference clip or click to upload

Supports WAV, MP3, M4A, OGG up to 30MB

Quick Starts

🔌 API Access

Call CosyVoice via the Eigen AI unified generate endpoint with your API key.

🔑 API Keys

To use the API, you need an API key. Manage your API keys →

POST /api/v1/generate

CosyVoice Speech Generation

Provide text, a reference script, and optional prompt audio to clone instantly.

Cost: 0.1 credits per audio minute

Minutes are estimated from output length. Upload reference audio for voice cloning or leave empty for default voice.

Request (cURL)

curl -X POST https://app.eigenai.com/api/v1/generate \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -F "model=cosyvoice" \
  -F "text=Hey there! I am your friendly voice twin. Pick a preset or upload your voice!" \
  -F "ref_text=The device would work during the day as well if you took steps to either block direct sunlight or point it away from the sun." \
  # Replace the path below if you supply your own reference clip
  -F "prompt_audio=@/path/to/reference.mp3" \
  --output cosyvoice_output.wav

Request (Python)

from pathlib import Path
import requests

url = "https://app.eigenai.com/api/v1/generate"
headers = {"Authorization": "Bearer YOUR_API_KEY"}
data = {
    "model": "cosyvoice",
    "text": "Hey there! I can read anything you send my way.",
    "ref_text": "The device would work during the day as well if you took steps to either block direct sunlight or point it away from the sun."
}
prompt_path = Path("/path/to/reference.mp3")  # Update to your reference audio path

if prompt_path.exists():
    with prompt_path.open("rb") as prompt:
        files = {"prompt_audio": (prompt_path.name, prompt, "audio/mpeg")}
        response = requests.post(url, headers=headers, data=data, files=files)
else:
    response = requests.post(url, headers=headers, data=data)

if response.status_code == 200:
    with open("cosyvoice_output.wav", "wb") as wav:
        wav.write(response.content)
else:
    print("Error:", response.text)

Request (JavaScript/Node.js)

const fs = require("fs");
const FormData = require("form-data");
const axios = require("axios");

const form = new FormData();
form.append("model", "cosyvoice");
form.append("text", "CosyVoice is ready to narrate your next release notes.");
form.append("ref_text", "The device would work during the day as well if you took steps to either block direct sunlight or point it away from the sun.");
const promptPath = "/path/to/reference.mp3"; // Update to your reference audio path
if (fs.existsSync(promptPath)) {
    form.append("prompt_audio", fs.createReadStream(promptPath));
}

axios.post("https://app.eigenai.com/api/v1/generate", form, {
    headers: {
        ...form.getHeaders(),
        Authorization: "Bearer YOUR_API_KEY"
    },
    responseType: "arraybuffer"
}).then((response) => {
    fs.writeFileSync("cosyvoice_output.wav", response.data);
});

Parameters

Parameter	Type	Required	Description
`model`	string	✅ Yes	Must be `cosyvoice`.
`text`	string	✅ Yes	Content CosyVoice will narrate.
`ref_text`	string	✅ Yes	Sample line voiced by the reference clip.
`prompt_audio`	file	Optional	Reference audio clip. Defaults to Eigen AI voice if omitted.

Output Preview

🎧 Powered by CosyVoice

Waiting for your first request. Enter text, add an optional reference, and generate to listen.