Skip to content
New · the open voice benchmark is liveRead it
cantari
Platform

Who owns the audio you generate

You own your generations: commercial use, worldwide, no watermark, no attribution. What we store, what delete does, and your cloning responsibilities.

Updated June 11, 2026

You own what you make

The audio you generate is yours. The moment you generate and export a file, you can use it commercially, worldwide, with no watermark and no attribution required. There is no separate license to buy and no royalty owed. Exports are standard MP3 or WAV files; nothing proprietary, nothing locked to our player.

This is the product stance in plain words, and it is part of the terms of use. The full plain-language summary lives on the ownership page.

License termAnswer
Commercial useYes
Attribution requiredNo
ExportsMP3 + WAV
WatermarkNone
TerritoryWorldwide
Your libraryPrivate

One honest caveat: engine terms

We route your work to several engines, and each engine vendor sets its own underlying license text. The per-engine rights summaries on the ownership page and the engines page reflect our honest reading, but where a summary and a vendor term ever disagree, the vendor term governs that engine's output. In practice the engines we run today allow commercial use of your outputs.

Voice cloning requires a consent attestation before anything is built: you confirm the voice is yours, or that you have the speaker's permission. The attestation ends with "I understand cloned voices I create are mine to use and mine to answer for", and we mean it. The server refuses to clone without it.

The acceptable-use rules apply with extra weight here: no impersonating a real person, no implying someone said something they did not, no fraud. Accounts that break those rules can be suspended.

Ownership of a cloned voice's output does not launder consent. You are responsible for having the speaker's permission and for what the clone says.

What we store

Your scripts and generated audio are kept in your private library: a private storage bucket scoped to your account. Nothing in it is public; playback and downloads use short-lived signed URLs that expire after about an hour, generated only for you while you are signed in.

If you clone a voice, the reference clip is stored privately in the same way, attached to that voice. We do not publish your work and we do not use it to train anything.

What delete does

Delete any library item individually, at any time: the stored audio and its record are removed. Deleting a cloned voice removes the voice from your account, deletes it from the engine that holds it, and removes the stored reference clip.

Files you already exported are plain MP3 or WAV files on your machine, and they stay yours forever regardless of what you delete in the app. For full account deletion, email us and a real person will handle it.

Where Cantari stands

Cantari is an early product under active development. There is no uptime guarantee yet, and the terms say so plainly. Keep your own copies of exported files you depend on; since exports are standard files with full rights, that is easy to do.