Runs on your device · Your file never leaves your device.
Record a clip or drop any audio or video file, pick a voice, and hear the result in seconds: deep voice, chipmunk, giant, robot, metallic, radio, megaphone, echo chamber, slow motion — or an anonymous voice that keeps you unrecognisable in a voice note you would rather not be traced to. Two sliders then nudge pitch (in semitones) and speed independently, so a lower voice does not have to become a slower one.
The effect engine is a browser port of the open-source Voicebox studio’s effect chains — the same reverb (Freeverb), compressor, filters and flanger, with the same preset values — plus a pitch shifter that preserves timing. Everything runs in your phone or computer’s own processor; the recording is not uploaded, and once the page is open the audio effects work with the connection off. That matters here more than most places: a voice is biometric data.
It works on voice notes from WhatsApp and Telegram (OPUS, OGG, M4A), phone recordings, MP3s, and the audio track of any MP4 or MOV. The output is a normal MP3 or WAV you can send straight back into any chat.
Not for: Cloning or imitating a real person’s voice — this changes how you sound, it does not turn you into someone else.
The pitch slider stretches the recording in time and then plays it back faster or slower, which brings the length back to where it was. That moves everything in the voice by the same amount, including the resonances of the throat and mouth that tell a listener how big the speaker is. So −3 sounds like a deeper version of the same person, while −7 or +6 sounds like a much larger or smaller one. For a natural result, stay within three or four semitones.
Each voice sets its own starting point: Deep −3, Anonymous −4, Giant −7 at 0.95×, Chipmunk +6, Slow motion −2 at 0.7×. The others leave pitch and speed alone and change the sound with filters, compression, chorus, ring modulation, distortion or reverb. The sliders then move from that starting point, so Radio with the pitch at −3 gives a deep radio voice. Reset puts both sliders back to the chosen voice’s values.
Anonymous lowers the voice by four semitones, keeps only the band from about 200 to 4,500 Hz, adds a light chorus and evens out the loudness. It does not change your words, accent, rhythm or pauses, or the room and background sounds around you.
Because the setting is fixed, someone who knows it can shift the pitch back and undo much of the change. Moving the pitch slider to a value of your own makes the change harder to guess.
Preview 10 seconds processes only the first ten seconds of the file, so a recording that opens with a pause gives a quiet preview. A recording over five minutes is processed in mono to save memory, and the result says so. Recordings over thirty minutes are refused; cut out the part you need on /mp3-cutter first. A surround soundtrack is folded to stereo with its centre channel, where dialogue usually sits.
When the file has a picture, a third output, Video with the new voice, puts the changed sound under the original picture in an MP4, with the sound stored as AAC. The picture is copied as it is when MP4 can hold it, and re-encoded to H.264 when it cannot. For this output the speed stays at 1×, even for Giant and Slow motion, so the voice keeps the video’s length and stays on the lips.
This page in other languages: العربية · Deutsch · Français · Español · Русский · Português · 中文 · हिन्दी · Türkçe · עברית · فارسی · اردو · Bahasa Indonesia · বাংলা