Turn a script into an AI voice-over

Paste the script, pick one of the voices and the speed, and an open text-to-speech model (Kokoro, Apache 2.0) speaks it in your browser. Download the MP3 or WAV, or put it on your video: the video's own sound is lowered under the voice and the picture is copied without re-encoding.

Every step runs in your browser; your files are not uploaded.

Steps

  1. Make the voice-over

    Type or paste the script. Write numbers and abbreviations the way they should be spoken, and add [pause 2] where you want a two-second pause. Pick the voice, the speed and the pause between sentences, listen, then download.

    Open as a full tool: AI Voice Generator

  2. Trim it (optional)

    Cut a breath or a pause at the start or end. Select the part to keep and download it as MP3 or WAV.

    Open as a full tool: Audio Cutter

  3. Put it on a video (optional)

    Add your video. Choose whether its own sound is lowered under the voice (by 14 dB unless you change it), kept, or removed, and when the voice starts. The picture is copied as it is, so there is no quality loss.

    Workflow step

FAQ

Can I use the voice-over commercially?

The Kokoro model is released under the Apache 2.0 licence, which allows commercial use. Nothing is added to the audio.

Does it sound like ElevenLabs?

It is a small open model that runs on your device, so it is not the same: it is clear and natural for narration, with fewer voices and less emotion control.

Made by BufferMint. We also design and build software products for founders. See our services →

Runs in your browser. Your files are not uploaded. We count anonymous visits, without cookies.