AI Voice Cloning - Clone Any Voice

Clone any voice from 10 seconds of audio. Enter text in any language, adjust the speed, and generate natural-sounding speech. Voice model creation is free.

1 Select or create a voice model
My voices
Current voice:
Upload Voice Sample
10 seconds to 4 minutes — up to 20 MB
2 Enter Text & Settings Select or create a voice model first
0 / 5000
Type in any language. Your cloned voice will speak it.
Settings
Emotion
Speed 1.0x
Format
Results kept 7 days · tied to this browser — Sign in to access them across devices
Loading tasks…
Updates paused · Retrying

Voice Clone Samples

Children's Cartoon
Donald Trump
Original
Cloned
Little Engine Narration

And here comes the little blue engine, going up the hill. Tremendous hill, by the way — many people are saying it's the biggest hill anyone has ever seen. And the engine? The engine is going to make it to the top. Believe me.

Takeout Menu
Barack Obama
Original
Cloned
Takeout Menu Reading

Let me be clear. The number seven combo — the kung pao chicken with fried rice — represents outstanding value at twelve ninety-nine. Now, some folks might go with the lo mein. And that is their right. But I believe, deeply and sincerely, in the kung pao chicken.

Cat vs. Water Glass
Morgan Freeman
Original
Cloned
Cat Pushes Glass

She sat there, perfectly still, her amber eyes fixed upon the glass at the edge of the table. The universe held its breath. One paw reached out — slowly, deliberately — and with the quiet certainty of gravity itself, the glass tumbled to the floor. She did not flinch. She never does.

Office Lunch Documentary
David Attenborough
Original
Cloned
Office Lunch Safari

Here, in the fluorescent-lit plains of the open-plan office, the junior analyst approaches the communal microwave. He guards his leftover pasta with quiet determination. A rival appears — Karen, from Accounting — and the tension becomes palpable. Only one may reheat at a time.

How It Works

1
🎤

Upload Voice Sample

Upload a clear audio clip (10+ seconds) of the voice you want to clone. Supports MP3, WAV, M4A, OGG, and FLAC.

2
✍️

Enter Text & Settings

Type the text you want spoken in any language, then adjust the speed.

3
🔊

Generate & Download

Listen to the AI-generated speech and download as MP3 or WAV.

Frequently Asked Questions

Our AI analyzes the voice characteristics from your audio sample — tone, pitch, accent, and speaking style — then generates new speech that matches those characteristics using text-to-speech synthesis.

Yes. Record a short, clear clip of yourself speaking in a quiet room and upload it as MP3, WAV, M4A, OGG or FLAC. Create your voice from the sample at no cost, then type any script to generate speech in your own voice, with adjustable speed. Download the result as MP3 or WAV for videos, voiceovers, product demos or presentations.

You can start for free. Creating a voice from your sample costs no credits. Generating speech is billed by text length, from 1 credit, and a 500-character English script costs 3 credits. Signed-in free accounts get 10 free credits every day, enough to test the clone on short scripts before you buy credits. No subscription is required.

Creating a voice model is free; you pay only when you generate speech, and the cost scales with text length, starting from 1 credit. A 500-character English script costs 3 credits, while 500 Chinese characters cost 7, because each Chinese character is billed like three English letters. Choose MP3 or WAV before generating; each generation is charged only once. A single generation never costs more than 68 credits, even at the full 5,000-character limit. Signed-in free accounts get 10 free credits every day.

The sample must be 10 seconds to 4 minutes long, in a file of up to 20 MB. Within that range, a longer sample of clear speech generally produces a more accurate clone than a short one. For best results, record a single speaker in a quiet room, with minimal background noise and no music.

No language setting needed — type your script in any language and the cloned voice speaks it. The language of your reference audio does not limit the output: a voice cloned from an English sample can read Chinese, Japanese, Spanish, and more. Widely spoken languages sound the most natural.

Cloning your own voice or voices with explicit consent is legal in most jurisdictions. Never clone someone's voice without their permission. Check your local regulations regarding AI-generated audio.

Yes. Generated audio can be used commercially, provided you have permission from the voice owner. No attribution is required on paid plans or credit packs; free-plan outputs require attribution to LubanAI.

LubanAI charges per generation, with no subscription required. Creating a voice model is free, and speech generation starts from 1 credit, priced by text length. Signed-in free accounts get 10 free credits every day to test a cloned voice, and credit pack credits never expire. Results download as MP3 or WAV, and you can type your script in any language with no language setting.

Tasks are processed in the order they are submitted. Paid tasks from subscribers and credit pack buyers are picked up first. Free tasks may wait longer during peak hours.

We use cookies to improve your experience. See our Privacy Policy.