Fix a recording without re-recording
Regenerate one flubbed line of your podcast or video in your own voice.
New
AI Voice Cloner
Clone a voice from a 10 to 30 second sample and make it say anything, in 10 languages.
AI Voice Cloner
Press play to hear a real result.
Saved to your workspace history.
The RegiAI AI voice cloner copies a voice from a short sample and uses it to read new text. Upload 10 to 30 seconds of clear speech, type up to 2,000 characters, and in about 10 seconds you get a WAV file of that voice saying your words. It supports English, Spanish, French, German, Italian, Portuguese, Russian, Japanese, Korean and Chinese.
It costs $0.10 per 1,000 characters, so any text up to 1,000 characters counts toward your daily free generation. You may only clone your own voice or a voice you have permission to use. For stock voices with no sample, try the AI Voice Generator; to keep a recording and change who it sounds like, use the AI Voice Changer.
How to use AI Voice Cloner
-
1
Upload a voice sample
Add 10 to 30 seconds of one person speaking clearly, with no music. MP3, WAV, M4A or OGG up to 10 MB.
-
2
Type what it should say
Enter up to 2,000 characters, pick the language or leave it on automatic, and confirm you have permission to use the voice.
-
3
Clone and download
Press Clone and speak. The new speech is ready in about 10 seconds as a WAV file.
What you can make
Consistent channel narration
Keep your voice on every video, even on days you cannot record.
Speak other languages
Have your cloned voice read a script in Spanish, Japanese or eight other languages.
Courses and presentations
Narrate lessons and slides in the instructor's own voice from written notes.
Personal audiobooks
Record a story for your kids or family in your own voice without reading it all aloud.
Approved voice talent
Produce extra lines from a voice actor who has agreed to be cloned for the project.
Accessibility for your own voice
Keep a record of your voice and use it to read messages aloud if speaking becomes difficult.
Personalized greetings
Send birthday or holiday messages in your own voice, written instead of recorded.
How to record a sample that clones well
The sample is the single biggest factor in quality. The cloner copies whatever it hears, including room echo, hiss and background noise, so a clean sample gives a clean clone.
- Length: 10 to 30 seconds. Shorter samples miss the character of the voice; longer ones are not needed.
- One speaker only. Cut out any other voices, laughter or interruptions.
- No music or effects. If your only recording has music under it, run it through the AI Vocal Remover first and keep the vocals.
- Natural delivery. Speak the way you want the clone to sound. A flat, read-aloud sample gives a flat clone.
- Add the transcript. Typing what is said in the sample (up to 500 characters) helps the model align the voice and usually improves accuracy.
A phone voice memo recorded in a quiet room is usually enough. Hold the phone about a hand's width from your mouth and avoid rooms with hard, echoing walls.
Consent and responsible use
Voice cloning is powerful, so the tool requires you to confirm that the voice is yours or that you have permission to clone it. That box must be ticked before anything is generated.
Do not clone a voice to deceive, impersonate, harass or defraud anyone, and do not use a clone to put words in a real person's mouth without their agreement. That includes public figures. If you are making content with someone else's cloned voice, get their permission in writing and tell your audience when it matters.
Your cloned audio is saved in My creations on your dashboard for 30 days, and you can delete it at any time.
A practical rule: if you would feel uncomfortable showing the result to the person whose voice it is, do not make it. Cloning your own voice to save recording time, or cloning a colleague who has agreed to narrate a training course, are good uses. Pranks, fake endorsements and messages that pretend to come from someone are not.
Voice cloner, voice changer or voice generator?
These three tools solve different problems:
- AI Voice Cloner (this page): you have a voice sample and new text. Output is new speech in that voice. $0.10 per 1,000 characters.
- AI Voice Changer: you have a recording and want it to sound like another voice, keeping the words and timing. $0.05 per recording.
- AI Voice Generator: no sample at all; choose from 16 ready-made voices with emotions. $0.15 per 1,000 characters.
To turn a video into another language while keeping the speaker's voice, the AI Audio Translator does the translation and the voice matching in one step.
Many projects combine them. A common workflow is to record a short, clean sample of your own voice once, keep it on your computer, and reuse it with the cloner whenever you need new lines. Each clone is a new generation, so you can reuse the same sample as often as you like and pay only for the text you generate.
If the words should match an existing video, write the script to the same length as the original line. Cloned speech follows the natural pace of the speaker in your sample, so a similar length of text gives a similar duration.
AI Voice Cloner FAQ
Still stuck? Contact us.
Is the AI voice cloner free?
Signed-in users get one free generation a day on tools priced $0.10 or less. A clone of up to 1,000 characters costs $0.10, so it qualifies. Longer texts and extra clones are paid per use. The free generation does not stack or roll over.
How much does voice cloning cost?
$0.10 per 1,000 characters, rounded up, with a $0.10 minimum. The 2,000-character maximum costs $0.20. See pricing for credit packs.
How long should the voice sample be?
10 to 30 seconds of clear speech from one person. The file can be MP3, WAV, M4A or OGG up to 10 MB.
Can I clone my voice in another language?
Yes. Pick one of 10 languages (English, Spanish, French, German, Italian, Portuguese, Russian, Japanese, Korean or Chinese) or let it detect the language from your text. The accent follows the original speaker, so results vary by language.
Can I clone a celebrity or someone else's voice?
Only with their permission. The tool requires you to confirm that the voice is yours or that you are allowed to clone it. Never use a cloned voice to deceive or impersonate anyone.
Why does my clone sound robotic or noisy?
Usually the sample has background noise, music, echo or more than one speaker. Record a new 15 to 20 second clip in a quiet room and add its transcript.
What format is the output?
A WAV file, which keeps full audio quality. Import it directly into any audio or video editor, or convert it to MP3 if you need a smaller file for sharing.
Does it add a watermark?
No. RegiAI does not add a watermark or audio tag to cloned speech. That is also why consent matters: please use cloned voices openly and responsibly.
Can I use the cloned voice commercially?
You can use results under our Terms of Service and the model provider's license, and only if you have the voice owner's permission.
What if the cloning fails?
The charge is refunded to your balance automatically, or your free generation is returned for the day. Try again with a cleaner sample if it keeps failing.
Can I reuse the same voice sample again later?
Yes. Keep your sample file and upload it again with new text each time. Every clone is a separate generation, priced by the length of the new text.
Does the reference text have to be exact?
It should match what is said in the sample as closely as you can manage. It is optional, but an accurate transcript usually improves how closely the clone matches the voice.
Can I clone a singing voice?
The cloner is designed for speech. Singing samples and sung output are not supported well, so use spoken samples and spoken text.
Is there an API for voice cloning?
Yes. The voice cloner is available through the RegiAI API at the same price.