Clone your voice with ElevenLabs. In under 10 minutes.

Instant vs. Professional Voice Cloning

ElevenLabs offers two kinds of voice cloning. Instant Voice Cloning (IVC) works from roughly one minute of audio and produces a usable clone in minutes — ideal for prototypes, personal projects, and most creator workflows. Professional Voice Cloning (PVC) fine-tunes a dedicated model on 30 minutes to several hours of audio for near-indistinguishable results, at the cost of longer training time.

This guide covers IVC. Start there — if the result isn't close enough for your use case, graduate to PVC with the same recordings plus more material.

Step 1 — What you need

• 1.An ElevenLabs account on a plan that includes Instant Voice Cloning (the free tier lets you explore text to speech; cloning unlocks on the Starter plan and up).
• 2.1–3 minutes of clean audio of the voice you want to clone — your own voice, or one you have explicit written consent to use.
• 3.For the API route: an API key (created in your ElevenLabs dashboard under Developers → API Keys) and Python or Node.js installed.

Step 2 — Record audio the clone will thank you for

Clone quality is determined almost entirely by input quality. Before you record:

• Quiet room, no echo. Soft furnishings beat bare walls. No background music, fans, or traffic.
• One speaker only. Crosstalk or a second voice in the sample will contaminate the clone.
• Speak the way you want the clone to speak. The clone mirrors the energy, pace, and tone of the sample. Record conversationally if you want conversational output.
• Aim for 1–3 minutes total. More clean audio helps; more noisy audio hurts. Several short clips are fine — the API accepts multiple files.
• Use decent gear if you have it. A USB mic beats laptop mics, but a phone in a quiet closet beats a great mic in a noisy room.

Step 3 — Clone via the dashboard (no code)

1. Sign in to ElevenLabs and open Voices from the sidebar.
2. Click Add a new voice → Instant Voice Clone.
3. Upload your audio clips (or record directly in the browser).
4. Name the voice, add labels/description if you like.
5. Confirm you have the rights and consent to clone this voice, then create it.
6. Open Text to Speech, select your new voice, type a sentence, and generate. That's your clone.

Done in a few minutes. If you only need voiceovers in the studio, you can stop here. Building it into a product or pipeline? Keep going.

Step 4 — Clone via the API (Python or TypeScript)

Install the official SDK (pip install elevenlabs python-dotenv or npm install @elevenlabs/elevenlabs-js dotenv), put your API key in a .env file as ELEVENLABS_API_KEY — never hardcode it — and create the clone with voices.ivc.create:

# example.py
import os
from dotenv import load_dotenv
from elevenlabs.client import ElevenLabs
from io import BytesIO

load_dotenv()

elevenlabs = ElevenLabs(
  api_key=os.getenv("ELEVENLABS_API_KEY"),
)

voice = elevenlabs.voices.ivc.create(
    name="My Voice Clone",
    # Replace with the paths to your audio files.
    # The more files you add, the better the clone will be.
    files=[BytesIO(open("/path/to/your/audio/file.mp3", "rb").read())]
)

print(voice.voice_id)

// example.mts
import { ElevenLabsClient } from "@elevenlabs/elevenlabs-js";
import "dotenv/config";
import fs from "node:fs";

const elevenlabs = new ElevenLabsClient();

const voice = await elevenlabs.voices.ivc.create({
    name: "My Voice Clone",
    // Replace with the paths to your audio files.
    // The more files you add, the better the clone will be.
    files: [
        fs.createReadStream(
            "/path/to/your/audio/file.mp3",
        ),
    ],
});

console.log(voice.voiceId);

Run it with python example.py or npx tsx example.mts. The script prints a voice ID— save it. That ID is how you reference your cloned voice in every subsequent text-to-speech call, the same way you'd use any stock voice.

Step 5 — Put your clone to work

• Audio versions of written content — turn newsletters, blog posts, and docs into narrated audio in your voice.
• Video voiceovers — script changes no longer mean re-recording; regenerate the line.
• Multilingual content— your cloned voice can speak languages you don't, for dubbing and localization.
• Apps and agents — pair the voice ID with the text-to-speech or streaming API to ship voice features in your product.

Troubleshooting

Consent and ethics — not optional

Only clone voices you own or have explicit, documented permission to clone. ElevenLabs requires you to confirm you have the necessary rights, and misuse — impersonation, fraud, cloning public figures — violates their terms and, in many places, the law. If you publish AI-narrated content, disclosing that it's AI-generated builds trust rather than costing it.