Sign in Start free

Guides

What you may legally do with AI-generated audio

Two licences apply to every file you render. Most tools mention one.

Last checked 2026-08-14

In short

On Naadly, audio you render may be used commercially on every plan, including free. Separately, the open voice models carry their own licences - mostly CC BY 4.0, some Apache-2.0 - and a few ask for attribution, which is printed on each voice's page with the exact credit line. Cloning a real voice additionally requires the documented consent Naadly records for you.

The service licence

What you render is yours to use, including commercially, including in client work, on every plan. There is no separate commercial upgrade, no per-broadcast fee and no clause that reclaims the audio if you cancel - files you already rendered stay usable. What you may not do is listed in the acceptable use policy, and it is the obvious list: impersonation, deception, harassment, and reselling the raw voices as a competing voice library.

The model licences

The catalogue is built from openly licensed models, which is what makes commercial use possible without a per-character bill. Each carries its own terms - predominantly CC BY 4.0 with some Apache-2.0 - and CC BY requires credit. Naadly prints the licence and a copy-ready attribution line on every voice page and collects them all on the licences page, so a client asking "what is this voice" gets an answer rather than a shrug.

Voices whose licences forbid commercial use are not listed at all. That is why the catalogue is the size it is rather than the size it could be.

Ownership, and the honest caveat

Whether purely machine-generated output attracts copyright at all differs by country, and in several - including the United States for works without human authorship - it may not. This affects whether you can stop somebody else from using an identical file, not whether you may use yours. If that matters to your business, get advice for your jurisdiction; nothing on this page is legal advice.

Cloned and converted voices

A cloned voice adds a third party: the person whose voice it is. Naadly will not create one without a recorded consent statement, and keeps that record as evidence. Converting your own performance into a catalogue voice needs no extra permission; converting it into an impersonation of a real person is prohibited whatever the licence says.

Watermarking and disclosure

Every render carries an inaudible AudioSeal watermark and is served with a header declaring it synthetic, and a provenance record is kept. This is deliberately not optional: it is what lets you answer a broadcaster, a platform's disclosure field or a procurement questionnaire, and it does not affect what the audio sounds like.

Questions

Can I use AI voice over commercially?
On Naadly, yes, on every plan including free. Check the voice's page for an attribution requirement.
Do I have to credit the voice?
For CC BY voices, yes - the exact line is on the voice's page. Apache-2.0 voices do not require on-screen credit.
Do I own the audio?
You may use it freely and commercially. Whether it is copyrightable at all depends on your country's law on machine-generated works.
Is the audio watermarked?
Yes, inaudibly, with AudioSeal, plus a synthetic-media header and a provenance record. Removing them is against the terms.