Voice Cloning

You upload a short, clean reference sample and the tool clones that voice, then reads any text you type back in it. The quality of the clone is decided almost entirely by that one sample, so a clean recording matters far more than a long one.

A clear 10–20 second clip works best — long enough to capture natural rhythm and intonation, without needing more. It should be mono, at least 24kHz, and can be up to 60 seconds or 10 MB, but a short, clean sample beats a long, noisy one every time.

Both the cloning and the synthesis are pay-per-use through redeem codes, with no account required. You only pay for what you generate, so trying a sample is inexpensive.

WAV, MP3 or M4A files work, ideally mono and at least 24kHz, with 10–20 seconds of clear speech (up to 60 seconds and 10 MB). One person talking normally in a quiet room, with no music or background noise, gives the most natural clone.

Almost always the sample is to blame rather than the tool. Background noise, room echo or clipping from recording too loud all carry straight into the clone. Re-record a cleaner clip in a small, soft space and the result usually improves dramatically.

Only your own voice, or someone who has clearly agreed to it. Cloning a person's voice without their consent isn't permitted and can cause real problems — that's not what the tool is for.

With a clean 10–20 second sample the clone closely matches the original's tone and delivery. The real test is typing a sentence the sample never contained and listening — new words reveal the true quality, since any tool can echo back a phrase it was given.