Senteri reports on how machines read the web.
As AI agents become the first reader of most pages, the rules of what makes a site legible are being rewritten — and almost nobody is testing them in the open. This is where that testing gets published.
// the premise
For thirty years, "who reads your website" had one answer: a person, or a search crawler indexing it for a person. That's no longer true. A growing share of what arrives at a page now is an AI agent — reading on someone's behalf, comparing options, filling forms, completing actions.
Agents don't see your layout. They read your HTML, your structured data, your text — and act on what they find. That changes what it means for a site to be legible, trustworthy, and safe. Senteri covers that shift: agent-readiness, the protocols underneath it, and what happens when your page's first reader isn't human.
// how it's reported
The strongest format here is the experiment: a real test against a live page or a real agent, with the result reported straight — including the failures, which are usually the most useful part. Alongside the experiments are essays that work out the ideas, and field reports tied to the six layers of agent-readiness.
Nothing here is a prediction dressed as a finding. If something was tested, it says what was tested. If something is still unsettled — and a lot is — it says so.
// where it comes from
Senteri is the English-language front of a larger, Polish-language research engine that has been working this beat daily since well before it had an English name. That engine ships daily threat intelligence, maintains a 280+ term agentic-web dictionary, and runs a working MCP layer. Senteri selects from it, rewrites it, and sharpens it for a global audience — it isn't a translation, it's the stronger version.