What RSL Is
RSL, or Really Simple Licensing, is a machine readable file that states the terms on which AI systems may use your content. It is the third file in a small stack that has been forming for two years, sitting alongside robots.txt and llms.txt, and it answers a question the other two do not.
The Three Files and What Each Answers
These get confused constantly, and the confusion produces files that contradict each other. They answer three separate questions.
| File | Question it answers | Enforced by |
|---|---|---|
| robots.txt | May you read this? | Crawler compliance, widely honoured |
| llms.txt | What is this site and what matters on it? | Nothing, it is a summary for models |
| RSL | On what terms may you use what you read? | Nothing technical, it is a licensing declaration |
Notice that only the first one controls anything. The other two are statements, and a statement is still useful, but expecting enforcement from them is the mistake that makes people dismiss the whole idea.
What RSL Actually Declares
Permitted uses
Whether content may be used for model training, for retrieval and citation, for commercial products, or only for personal reference. These are separable and most publishers want different answers for each.
Conditions
What an operator must do in exchange, most commonly attribution, and in some cases a payment or agreement reference.
Scope
Which parts of the site the terms cover, because a documentation section and a paid research archive rarely deserve the same answer.
It States Terms, It Does Not Enforce Them
This is the part to be honest about. Publishing RSL does not stop anybody from using your content, in the same way a copyright notice does not prevent copying. It is a declaration, not a technical control.
What it is actually good for
Making your terms explicit and machine readable. That matters for operators that check, and it matters in any future dispute where what you intended is relevant. It does not matter at all if you were never going to enforce anything.
Who Should Publish One
Publish RSL if your content is the product. Publishers, data providers, paid research, documentation sold under licence, anything where reuse without terms is a direct revenue loss. For these, stating terms costs nothing and closes an ambiguity that currently defaults against you.
Who Can Safely Skip It
Skip it if your content markets a product. A software company whose blog exists to attract customers wants the widest possible reuse, because every quotation is free distribution. Publishing restrictive terms on marketing content is working against your own objective, and publishing permissive terms is a formality that changes nothing.
That covers most businesses reading this. The honest answer for a typical SaaS or services site in 2026 is that RSL is optional, and the effort is better spent on being worth quoting in the first place.
Where the File Goes
At a predictable location on your domain, following the same convention robots.txt and llms.txt use. The entire value of a well known location is that an agent can find it without being told, so putting it somewhere custom removes most of the benefit and all of the discoverability.