Susurration Observatory

Susurration Observatory measures how AI agents get on with websites: whether they can find what they are asked to find, and where they get stuck.

How to recognise it

Our browser sends a normal Chromium user agent with this token added:

Susurration-Observatory/<version> (+https://susurration.ai/observatory)

We never pretend to be another crawler or a person, and we do not try to get around bot protection. If a site blocks us, we record that and move on.

It reads pages the way a visitor would: opening pages, using search, reading policies and product pages.
It never sends a form, places an order, logs in, creates an account or pays. Our browser blocks every request that would change something on the site, apart from adding an item to a session basket where a test asks for it.
One test at a time, with at least a minute between tests on the same site.

Only sites whose owner has agreed, and our own sites. That will widen later; this page will say so first.

We follow robots.txt for the token Susurration-Observatory. To keep us out:

User-agent: Susurration-Observatory
Disallow: /

Questions or a request to stop: info@hiveminds.nl, with "Observatory" in the subject.