Make any website agent-ready — and keep it that way

Sichta Conformance checks whether AI agents can read your site, generates the files that fix what fails, and tells you when it drifts.

Deterministic pass/fail checks. No visibility promises — just facts.

Check, fix, verify, watch

  1. 1

    Check

    Seventeen deterministic checks across four pillars. Every failure has a typed error, an explanation and a fix — never a score you have to interpret.

  2. 2

    Fix

    We generate llms.txt, llms-full.txt, a JSON-LD snippet and security.txt from your own pages. Where a page has no description, you get a marker to fill in — we never invent one.

  3. 3

    Verify

    Confirm the files are live and a verification check runs within five minutes. You get the result by email.

  4. 4

    Watch

    A daily check. If something that passed starts failing, you hear about it once — not every day.

What gets checked

  • Discovery surface

    llms.txt exists, parses, and its links resolve. Sitemap present. Nothing blocked by robots.txt.

  • Access policy

    robots.txt parses. AI crawlers are treated consistently. Your homepage does not refuse them at the CDN.

  • Structured data

    JSON-LD present and valid. An organisation entity with a name and URL. sameAs links that resolve.

  • Machine endpoints

    /.well-known/ reachable. A UCP manifest, if you publish one, is valid. Markdown content negotiation answered honestly.

What this does not do

  • It does not promise visibility, rankings or traffic. Nobody can, and a tool that says otherwise is selling you a feeling.
  • It does not guess. Where a check cannot be determined, it says so and does not count it against your score.
  • It does not write your descriptions. A generated sentence about your business, indistinguishable from one you wrote, is not a fix.