Is my recording uploaded?

No. Your files are never uploaded: the recording is made, changed and saved by your own browser, on your device.

Which model does this use?

None. Pitch and formant shifting is Signalsmith Stretch, a 113 KB audio engine that runs in your browser, and robot, radio and echo are built from the browser's standard audio filters. Because there is no AI voice model, there is nothing extra to download.

Can I use it live in Discord, a game or a call?

No. Other programs hear you through a microphone, and a web page cannot add a microphone to your computer — that takes an installed app with a virtual audio device. What this does is change a recording: make the clip here, then send it or post it wherever it is going.

Can it make me sound like a celebrity or a particular person?

No, and that is deliberate. The effects change your own voice — higher, deeper, bigger, smaller, robotic — but none of them imitates anyone else.

Can it make a male voice sound female, or the other way round?

Higher is the closest. It raises the pitch and the formants together, the two things that most change how big and how old someone sounds, and deep does the reverse. Neither fully convinces on its own, because the way someone talks carries a lot too. If the result sounds cartoonish, bring the formant fader back towards zero.

Will an effect stop people recognising me?

It depends on the effect. We tested each one with a speaker-recognition model: chipmunk and monster scored like a different person, deep and higher moved a long way towards one, and robot, radio and echo left the voice easy to recognise. No effect changes your accent, your words or the way you speak, so people who know you may still recognise you.

Why is live listening only on computers, and why the delay?

To change pitch, the effect has to hear a short slice of your voice before it can play it back, about 40 ms, and your computer's audio adds its own delay on top. On the computer we measured, that came to about a tenth of a second; robot, radio and echo skip the pitch step, so they come back sooner. Phones add more delay, Bluetooth headphones much more, and a phone's speaker sits right next to its microphone, so on phones the page offers recording instead. The delay never affects the file you download.

Can I use a video file?

Yes. Drop the video in and the sound is taken from it. What you get back is audio only, as an MP3 or a WAV.

What is the difference between the MP3 and the WAV?

The MP3 is 320 kbps and small enough to send. The WAV is lossless 24-bit, which is what you want if the file is going into an editor. If you only want to share it, take the MP3.

Change your voice: deep, high, robot or chipmunk

Record yourself or drop in a clip, pick an effect, and download the result. Seven effects cover what people usually want — deep, higher, chipmunk, monster, robot, radio and echo — with pitch and formant faders underneath if you want to go further. It is plain audio processing in your browser, not an AI voice model, so a minute of audio is ready in a few seconds.

Den här sidan finns också på svenska. Visa på svenska

Diese Seite gibt es auch auf Deutsch. Auf Deutsch ansehen

Esta página también está en español. Ver en español

Cette page existe aussi en français. Voir en français

Essa página também está em português. Ver em português

Questa pagina esiste anche in italiano. Passa all'italiano

Your voice
Drop a recording or a video here
MP3WAVM4AMP4
Your files are never uploaded.
Pick an effect
Your file

How to change your voice

Record or drop a clip

Press record and talk, or drop in an audio or video file. MP3, WAV, M4A and phone voice memos all work.
Pick an effect

Choose one of the seven effects and press play to hear it on your own recording. Move the pitch and formant faders to fine-tune it, or switch to the original to compare.
Download it

Make the file and take it as an MP3 to send, or as a lossless WAV to edit further. It comes back exactly as long as the clip you put in.

What you get

Seven effects, one click each

Deep and higher make you sound like a different person. Chipmunk and monster push it into a cartoon. Robot, radio and echo change the sound around your voice. You hear each one on your own recording before you make a file.

Pitch and formant, separately

Pitch is how high your voice is. Formants are the resonances that tell a listener how big the person talking is. Moving the two separately is why the deep effect sounds like a bigger person rather than a slowed-down tape, and the two faders let you find your own balance.

Hear yourself live

On a computer, switch on live listening and talk: the effect comes back through your speakers about a tenth of a second later. Wear headphones, or your speakers will feed back into the microphone. This is a preview on this page, not a microphone for Discord or games.

No AI model, no voices borrowed

The pitch shifting is Signalsmith Stretch, a 113 KB audio engine that loads with the page, and the other effects are the browser's own audio filters. There are no celebrity or character voices: every effect changes yours, and none of them copies anyone else's.

Frequently asked questions

Is my recording uploaded?: No. Your files are never uploaded: the recording is made, changed and saved by your own browser, on your device.
Which model does this use?: None. Pitch and formant shifting is Signalsmith Stretch, a 113 KB audio engine that runs in your browser, and robot, radio and echo are built from the browser's standard audio filters. Because there is no AI voice model, there is nothing extra to download.
Can I use it live in Discord, a game or a call?: No. Other programs hear you through a microphone, and a web page cannot add a microphone to your computer — that takes an installed app with a virtual audio device. What this does is change a recording: make the clip here, then send it or post it wherever it is going.
Can it make me sound like a celebrity or a particular person?: No, and that is deliberate. The effects change your own voice — higher, deeper, bigger, smaller, robotic — but none of them imitates anyone else.
Can it make a male voice sound female, or the other way round?: Higher is the closest. It raises the pitch and the formants together, the two things that most change how big and how old someone sounds, and deep does the reverse. Neither fully convinces on its own, because the way someone talks carries a lot too. If the result sounds cartoonish, bring the formant fader back towards zero.
Will an effect stop people recognising me?: It depends on the effect. We tested each one with a speaker-recognition model: chipmunk and monster scored like a different person, deep and higher moved a long way towards one, and robot, radio and echo left the voice easy to recognise. No effect changes your accent, your words or the way you speak, so people who know you may still recognise you.
Why is live listening only on computers, and why the delay?: To change pitch, the effect has to hear a short slice of your voice before it can play it back, about 40 ms, and your computer's audio adds its own delay on top. On the computer we measured, that came to about a tenth of a second; robot, radio and echo skip the pitch step, so they come back sooner. Phones add more delay, Bluetooth headphones much more, and a phone's speaker sits right next to its microphone, so on phones the page offers recording instead. The delay never affects the file you download.
Can I use a video file?: Yes. Drop the video in and the sound is taken from it. What you get back is audio only, as an MP3 or a WAV.
What is the difference between the MP3 and the WAV?: The MP3 is 320 kbps and small enough to send. The WAV is lossless 24-bit, which is what you want if the file is going into an editor. If you only want to share it, take the MP3.

Your voice

Pick an effect

Your file