Rights and disclosure when you adapt someone else's novel
Who owns the adaptation, what an option actually grants, why the pipeline cannot check copyright, what AI disclosure platforms require in 2026, and the four questions to settle before you import a single chapter.
The pipeline will adapt anything you paste into it. That is a technical fact, not a permission. This article is about the permission part, because it is the part that ends channels.
Who owns what
You keep the copyright in your manuscript. You also keep the copyright in the drama you produce from it — the video, the audio, the subtitles. We claim no licence over either, and we do not train models on your text unless you explicitly opt in.
That is simple when you wrote the novel. It gets complicated in three specific ways.
Three situations that need a written answer
1. You are adapting a novel you did not write. You need an adaptation right, not a reading right. Buying the ebook, or having the author’s public blessing, is not a licence. What you need in writing: the right to adapt to audiovisual, the territories, the term, exclusivity, and whether the author shares in revenue. Web-novel platforms sell these as standard option agreements; independent authors usually do not have one drafted, which means you draft it.
2. You are adapting a novel published on a platform. Many platform terms grant the platform audiovisual rights, which means the author may not be able to license it to you even if they want to. Check the platform’s author agreement before you sign anything with the author.
3. You are adapting a novel that is in the public domain. Public domain is per-territory and per-translation. A 1920s novel may be free in the US and still protected in the EU; a modern translation of a public-domain work is itself copyrighted. Adapt from the original language where you can.
What the pipeline cannot do
It cannot tell whether you hold the rights. It does not run a plagiarism check against published works. It will produce a clean, professional season from a pirated file, and the resulting video is not a defence against an infringement claim — “the AI made it” has not worked in any jurisdiction we are aware of.
A takedown lands on the account that imported the text and on the channel that published it. The platform that published the source is not a party.
Disclosure: what platforms actually require
As of late 2026, the major short-video platforms require creators to label realistic synthetic media. In practice:
- Masters carry a content-credential metadata block identifying the content as synthetic.
- You still tick the platform’s own disclosure box at upload — metadata is not a substitute.
- Removing the credential block to pass off generated video as live action is a terms violation on our side and, on several platforms, grounds for removal.
Likeness and voice
Generated cast members are synthetic. Uploading a photo or a voice sample of a real person to clone their likeness is prohibited, including for internal tests. This is the one rule where “we were only testing” has never been accepted as a defence, and it is also the fastest way to get a payment processor to drop you.
Four questions to settle before import
- Do I hold written adaptation rights for this text, in these territories, for this term?
- Does a platform’s author agreement already control the audiovisual rights?
- Is the translation I am working from itself licensed?
- Does my cast include anyone resembling a real, identifiable person?
Four yes/no answers, ten minutes. Cheaper than a takedown.