AI Terms / what it is
What is AI Terms?
AI Terms is a file a website publishes at /.well-known/ai-terms that says, for each use of its content by AI (search, answer, training, agent), whether that use is allowed, allowed for a price, or not allowed.
In one example
A site that lets AI search engines, AI answers and agents in, and says no to training:
That is the whole idea. Written as a file, it looks like this (default is the answer for a visitor that doesn't say what it is doing):
Code20 lines
{
"ai_terms": "0.1",
"site": "https://example.com",
"updated": "2026-11-16",
"default": "deny",
"uses": {
"search": {
"access": "allow"
},
"answer": {
"access": "allow"
},
"training": {
"access": "deny"
},
"agent": {
"access": "allow"
}
}
}Swap "allow" for "paid" with a price, and the same file tells an agent what a page costs and how to pay for it.
The four uses
How it sits beside the other files
| File | What it says | Per use? | Price? |
|---|---|---|---|
| robots.txt | Which crawler names may fetch which paths | No | No |
| llms.txt | Which pages an AI should read first | No | No |
| RSL license | Licensing terms for content, as an XML document | Yes | Yes, through a license server |
| AI Terms | Allow, paid or deny for each of four uses | Yes | Yes, paid in the request itself (HTTP 402) |
These work together. AI Terms does not replace robots.txt, and a terms file can point to an RSL license.
Questions
How is AI Terms different from robots.txt?
robots.txt says which crawler names may fetch which paths. It cannot say what the content may be used for, or name a price. AI Terms adds the use and the price, and works alongside robots.txt; it does not replace it.
Is AI Terms a standard?
No. It is an open draft, published as an individual Internet-Draft at the IETF (draft-ovalles-ai-terms). No standards body has adopted it. Anyone may implement it.
Does a site need AIPayGen to publish AI Terms?
No. A site that only allows or denies uses can write the file by hand or with the free generator and host it itself. Charging a price needs something that can take an HTTP 402 payment; AIPayGen is one such service.
How does an AI agent use AI Terms?
Before reading a site, it fetches /.well-known/ai-terms, looks up the rule for what it is about to do, stays out if that use is denied, pays first if it is paid, and sends an AI-Purpose header saying which use it is.
Its status, plainly
AI Terms is a draft, version 0.1. It was written by AIPayGen, is published under an open license (CC BY 4.0), and is posted at the IETF as an individual Internet-Draft. It is not a standard and no standards body has adopted it. The format, a JSON Schema, test cases and a small reference reader are public so that anyone can implement it without us.
Publish yours
Make a file in a minute Check a file Read the spec
A site whose live file passes the check is added to the public list of sites that publish AI Terms.
Once your file is live, you can show this badge. It links here, so visitors can see what it means:
Code1 lines
<a href="https://aipaygen.com/what-is-ai-terms"><img src="https://aipaygen.com/ai-terms-badge.svg" alt="This site publishes AI Terms" height="20"></a>How to refer to it
The name is AI Terms, two words, both capitalised. The file is ai-terms, served at /.well-known/ai-terms. A site "publishes AI Terms"; an agent "reads AI Terms".