Terms of use

Updated 11 October 2026.

Gullible is a free, open-source research tool. By using it you accept these terms.

What it does

It gives you a fictional web page that hides instructions aimed at AI agents, and tells you which of them your agent followed. The people, publications, products and places on those pages are invented. Any likeness to a real one is a coincidence.

What you may do

• Test an agent that you own, or that you have permission to test.
• Share a result, with its summary and its limits.

What you may not do

• Give a trap page to someone else's agent, or to an agent working on a shared or business account, without the permission of whoever is responsible for it.
• Try to overload the service, bypass its limits, scan it, use its collector as a tracking pixel or a message relay, or inject your own content into its pages.
• Present a result as proof that an agent is safe, or that it is unsafe. A result describes one session at one moment. The methodology explains what it does not measure.

No warranty

The service and its results are provided as they are, with no promise that they are accurate, complete or always available. We may change or stop the service at any time. Sessions and shared results expire.

Liability

Nothing in these terms limits a liability that the law does not allow us to limit, or the rights you have as a consumer. Within that limit, we are not liable for indirect losses that come from using a free research service.

The code and the names

The source code is released under the Apache License 2.0 (https://github.com/inari-suite-org/gullible). The names Gullible and Inari Suite and the logo are not part of that licence.

Abuse and security reports

Write to inari-suite@proton.me. We answer reports about a page or a vulnerability as quickly as one person can.

Law

These terms follow French law. If you are a consumer, you keep the protection of the mandatory rules of the country where you live, and you may bring a claim before its courts.