A free AI voice changer that puts any video or recording in another voice and keeps every word, the timing and the delivery, in your browser.
5.0(1 rating)Made a video with an AI tool, in a voice that isn't yours? Put it in your own voice, or any other, and keep every word and the timing. It runs in this browser, and nothing is uploaded.
Your video or recording
The new voice
Change the voice
Hear it and save it
Your file in the new voice shows here, to hear before you save it.
Your file and your voice never leave this browser. The models download from our server once, and nothing you add is uploaded.
This free AI voice changer puts any video or recording in another voice, and keeps every word and the timing.
Made a video with an AI tool, in a stock voice? Add a few seconds of your own voice, and the video speaks in it.
It runs in your browser, on your own graphics card, so your file and your voice never leave your computer.
I built it next to my voice cloning tool. Here's what a voice changer is, how it works, what I measured, and how to use it the right way.
What is an AI voice changer?
An AI voice changer takes a recording of someone talking and gives it a different voice. The words, the pauses and the emotion stay the same; only the voice changes.
ElevenLabs calls its version Voice Changer, and it used to be called speech to speech. The idea is the same here: your performance, a new voice.
How the AI voice changer works
First, the same model my music remover runs takes the voice off the music and the background. Then the tool changes the voice in one of 2 ways, and puts it back over the untouched background.
These are the steps of Best likeness, the one it picks for you.
- It writes down each line – with the time each one starts and ends, on your device.
- It says each line in the new voice – from the new voice's sample and its words, the way my voice cloning tool does.
- It fits each line into its slot – a line that runs long is sped up a little, up to 1.5 times, with its pitch kept, so it ends before the next one starts.
Keep the delivery works another way. It turns the voice into speech codes that carry the words and the timing, then renders those codes in the new voice.
That keeps the original performance to the syllable: 24.2 seconds in, 24.6 out in my test. It sounds less like the new voice, so it's the one for a real person's performance, not an AI voice-over.
How to change the voice in a video
These are the steps, from your file to the saved video.
Your own voice from a link goes through the same voice check as my voice cloning tool: you read one sentence, and it has to match.
What I measured
The test that matters most: a 15.5-second AI voice-over, the kind an AI video tool gives you, put in my voice from 26 seconds of me presenting. A speaker recognition model scored each version from -1 to 1.
- Sounds like me
- 0.06
- Sounds like the AI voice
- 1.00
- Words right
- All
- Sounds like me
- 0.77
- Sounds like the AI voice
- 0.05
- Words right
- All
- Sounds like me
- 0.56
- Sounds like the AI voice
- 0.20
- Words right
- 1 word off
My real voice scores 0.81 to 0.86 against other clips of me, so Best likeness lands close to the real thing, and the AI voice is gone. It came out 15.5 seconds long, like the original, and took 68 seconds in all the first time, downloads included.
I also changed my own voice into a woman's, and another person's voice into mine, with Keep the delivery. A speaker recognition model scored each result from -1 to 1 against the new voice; a speech recognition model wrote the words back down.
- Words
- Every word right
- Sounds like the new voice
- 0.52 (my own voice: 0.30)
- Words
- Kept
- Sounds like the new voice
- 0.35
- Words
- Kept
- Sounds like the new voice
- 0.42 (their own voice: 0.17)
A woman's voice from a clean sample came through clearly, with every word right. My own voice was harder to reach: it got closer with more of me, 30 seconds beat 10.
The voice changer keeps the delivery, so it's never quite as close to the new voice as a clone saying fresh text. For the closest match to a voice, use voice cloning and type the words.
Voice changer or voice cloning?
They sound alike, but they start from different things. This table shows when to use each.
- Voice changer
- A recording of someone talking
- Voice cloning
- Text you type
- Voice changer
- The words, the timing and the delivery
- Voice cloning
- Only the voice
- Voice changer
- Videos, podcasts, a line said the wrong way
- Voice cloning
- New lines, voice-overs, other languages
Tips for the best result
A voice changer copies the performance, so the recording you start with matters as much as the new voice.
- Start with clean speech – one person talking, and the tool takes the music off for you.
- Give the new voice 20 seconds – the tool reads up to 20 seconds of it, and more of a voice gets closer to it.
- Match the energy – a calm voice sample in an excited video sounds off; pick a sample with the same energy.
- Keep it under 5 minutes – longer files go in pieces, so cut a long one into parts.
Who is the AI voice changer for?
It's for anyone with a recording in the wrong voice, or a voice they'd rather not use.
- Voice cloning – creators who want their own voice on a clip someone else recorded
- Video caption generator – video makers who change the voice, then caption the result
- Remove music from video – editors who only want the voice, or only the music
- Speech to text – anyone who needs the words of the new recording written down
It isn't for putting words in someone else's mouth: change your own voice, or use one you have permission for.
Use it the right way
A voice is part of who someone is. These keep a changed voice honest.
- Change into your own voice, or a voice you have permission for – the ready-made voices here were given freely to Mozilla's Common Voice.
- Say it's AI – when you publish a changed voice, tell your listeners.
Changing a voice is legal, but some uses aren't. Stay clear of these.
- Changing a recording into a real person's voice without their permission – laws like Tennessee's ELVIS Act protect a person's voice.
- Faking a call or a message – a changed voice used to fool someone is fraud.
Want a pro voice?
The best pro voice
AI voice changer FAQs
Questions about the ai voice changer? Here's what to know.
An AI voice changer takes a recording of someone talking and gives it a different voice.
The words, the timing and the emotion stay the same.
Drop the video into the tool, pick the new voice and click Change the voice.
Then save the video: the picture stays as it was, with the new voice over the same background.
Yes: click Add your voice, then record, upload or paste a link.
A voice from a link has to match you: you read one sentence, and it checks.
Yes, the result is as long as your file to within a fraction of a second.
That's why it lines up with the lips in a video.
Up to 5 minutes, the same as ElevenLabs' voice changer.
Cut a longer file into parts first.
No, your file and your voice stay in your browser.
Only the models download, once, from our server.
Not yet: it needs Chrome or Edge on a computer with a graphics card.
A version for phones is next.
Yes, it's free, with no account and no watermark.
I did.
I'm Navid Moazzez, and the AI voice changer is one of my free tools on navid.me.
Read more about me.
Navid.me is reader-supported. When you buy through links on this site, I may earn an affiliate commission. Learn more.
More free tools
The most actionable AI newsletter for founders
Every week, get proven AI strategies, curated tools, and step-by-step systems to grow your audience, create better content, and build a profitable creator business.
No fluff, no filler, no BS. Just five minutes each week that might level up your online business and life.
P.S. Sign up now to get free access to my ultimate AI tools guide for creators.





















