Keep one voice in the recording
Avoid dialogue, background television, music, and overlapping speakers. The recording should contain only your voice.
A familiar voice. A new script.
Free AI Voice Cloning Online
Make your own voice say something new. Start with a short recording, hear a generated example, then bring your scripts into a project.
Microphone demo without signup. Your own scripts with a free account.
A voice worth coming back to.
For the stories, characters, and ideas that sound like you.
Try it now · No signup required
Hear yourself say something new
Record a short sample and hear your own voice saying something you have never recorded. No signup required.
Real AI voice clone
Hear your own voice saying a new sentence.
Works in your browser
No app, no installation, no account.
Your audio is private
Recordings are deleted automatically after 24 hours.
More with a free account
Use your own text, save your voice, and get unlimited AltVoice Lite.
- 1Record
- 2AI clones
- 3Hear result
Click to start recording
Your microphone will be used in this demo.
Record with microphone
Available for guests
Save your voice
Free account required
Hear the reference. Then hear the clone.
Compare the source recording with a generated line. Listen for the voice’s character and pronunciation; the new script and delivery will not be identical to the reference.
Original
00:03
“This is a short sample of my voice. You can use it to clone my voice and generate new speech.”
Generated with AltVoice
00:06
“This is the same voice, generated from text. Natural, clear, and ready to use in your projects.”
How to clone your voice with your own text
The demo above reads a fixed sentence. Your workspace gives you control over the script and keeps the result in a project.
Record your voice
Create a free account and open a voice project. Record yourself with the microphone in a quiet room. AltVoice clones only your own voice, so audio files and other people's voices are not accepted.
Write a short test script
Start with a few sentences. Select an available speech model and, where supported, set an emotion or delivery direction.
Listen, refine, and save
Check the sound before generating more. Download the completed audio and keep the voice in your library to use for the next script.
Record a reference the model can work with
A clear voice recording is more useful than a dramatic performance. Try a clean 10–15 second passage in your normal speaking voice.
For example: “I’m recording a short voice sample in a quiet room. This is how I naturally speak when I’m telling someone a story.”
Keep the microphone steady
Speak at a comfortable volume and maintain a consistent distance. Move away from an echoing wall or a noisy fan rather than trying to speak over it.
Record in the style you need
A calm reference is a useful starting point for narration. A whispered or shouted reference can make a neutral line harder to produce.
Keep the voice. Change the story.
A creator can record a new introduction without setting up the microphone again. A narrator can test fresh lines in their own voice before a recording session. A storyteller can compare several readings before choosing the one that fits.
Cloning aims to preserve a voice’s character. It does not guarantee an identical performance, perfect pronunciation, or the same result in every language.
When the clone does not sound right
Change one part of the setup and compare the same short sentence again.
| What you hear | What to try next |
|---|---|
| The voice feels distant or noisy | Record in a quieter, less echoing space. Avoid music and heavily processed audio. |
| The identity is not close enough | Use a clearer passage from one speaker. Try another reference or an available model and compare both with the original. |
| A name or number sounds wrong | Write out the intended pronunciation or number in words. Test that line before generating the rest. |
| The rhythm feels unnatural | Split a long sentence, simplify punctuation, and remove conflicting delivery cues. |
| The accent changes in another language | Check the model’s language support and use a short test in the target language. Review it with a fluent speaker. |
Need a voice, but not a clone?
You do not need a recording for every project. Choose the starting point that matches what you want to make.
Ready-made voices
For a quick narration or explainer, pick a voice and enter your text. No microphone needed.
Start with text to speechOriginal character voices
Describe a voice that does not need to sound like a real person. Explore age, texture, tone, and delivery.
Create a voice from a descriptionAtmosphere and effects
Build the sound around your dialogue with scene descriptions: footsteps, machinery, weather, and more.
Generate sound effectsQuestions about cloning your voice
How do I clone my voice online?
Record a clear sample in the microphone demo and listen to a short generated example. For your own text, create a free account, open a voice project, record your voice with the microphone, then enter your script and generate.
Is free voice cloning available without signup?
The microphone demo works without signup and reads a fixed sentence. A free account lets you use your own text, save your recorded voice, save projects, and generate with AltVoice Lite.
How much audio should I record?
For the microphone demo, record 5–40 seconds. A clean 10–15 second passage is a useful starting point: speak naturally, with one speaker and no music. Saved voices in the workspace use 5–60 seconds.
Can I upload an existing recording?
No. AltVoice clones only your own voice, recorded live with your microphone in the browser. Audio files are not accepted as voice references, which is how we keep the service limited to voices people have actually authorized.
Why does my cloned voice sound different?
Background noise, reverberation, multiple speakers, and an exaggerated performance can change the result. Try a cleaner sample in your normal speaking voice, shorten the test script, and compare available models. A clone is a generated approximation, not a guarantee of an identical voice.
Can I make my voice speak another language?
With a model that supports your target language, you can use the same reference with a new script. Accent and pronunciation may change. Test a few sentences with a fluent speaker before publishing.
Can I save a voice and use it again?
Yes. Your account has a voice library and projects, so you can reuse a saved voice for new scripts without starting from the microphone demo each time.
How do I make the delivery more expressive?
Begin with punctuation and short, well-formed sentences. Where supported, use the model’s emotion controls. A reference spoken in the intended style can also help; avoid adding several conflicting instructions at once.
Can I clone another person’s voice?
No. AltVoice is built for cloning your own voice. Recording or playing back someone else’s voice into the microphone violates our Voice Consent Policy and Acceptable Use Policy and can lead to account termination. Use Voice Design if you need a voice that is not your own.
What happens to the recording from the demo?
The no-account microphone demo is temporary: its recordings and results are scheduled for automatic deletion after 24 hours. Account projects use a separate retention policy and can be managed from your dashboard.
Can I use my cloned voice commercially?
Check our commercial license and terms for your intended use. You need rights to your script; generating audio does not give you permission to use someone else’s identity or copyrighted material.
Voice consentData retentionCommercial licenseAsk us a question