How Kelvix reads your site.

The same way a search crawler would, and in the same order. Here is each step, what it produces, and where it stops.

1
We read your robots.txt first
Before anything else, because it is the file that decides what a crawler is allowed to touch, and because a rule in it is the single most common reason a page is invisible. We obey what it says, including any crawl delay, even when obeying it means we read less.
It is also where we read whether assistants are allowed in. ChatGPT, Perplexity and Claude each announce themselves with their own name, and a site can allow or refuse them separately.
What we read
```
User-agent: *
Disallow: /wp-admin/

User-agent: OAI-SearchBot
Allow: /

User-agent: PerplexityBot
Allow: /
```
2
We read up to 20 pages, one at a time
Starting at your home page and following the links it allows. One page at a time with a pause between, so a small site's hosting never notices we were there. It takes about a minute, and you watch it happen page by page rather than staring at a spinner.
Twenty is a deliberate ceiling, not a teaser. Nearly every problem worth reporting on a local business site is a template problem. It is on every page, and twenty pages is more than enough to prove that.
- read/200
- read/services200
- read/services/repair200
- read/about200
- read/contact301
3
You get findings, worst first
Each one says what it is, how many of your pages it touches, the evidence we actually read, and where the rule comes from. Sorted by what it costs you, so the first thing you read is the thing most worth your afternoon.
Severity is never colour alone. Every level carries its own icon shape and a written label, so the report survives being printed in black and white or read by somebody who cannot separate red from orange.
Example
CriticalPages blocked by robots.txt
1. What Kelvix found
  Your robots.txt file tells search engines not to visit this page at all.
  Where we found it
  /services
  /services/emergency-repair
  /book
2. How to fix it
  Why this matters, and how to fix it: what it costs you, the exact change to make, and where it goes.
  Show me the fix
Technical details · the evidence, all 3 pages, the source
What we found on the page
Disallow: /services Disallow: /book
The 3 affected pages
/services
/services/emergency-repair
/book
Based on published guidance — Google Search Central — AI features and your website

What we never do.

We never change your website. Kelvix only reads. It never submits a form, never signs in, never writes anything anywhere. Every fix is something you or your web person applies.
We never hide what we are. The crawler identifies itself as KelvixBot in every single request, pointing at a page explaining what it is — so a site owner checking their logs gets an answer rather than a mystery.
We never ignore robots.txt. Including our own entry: two lines in your robots.txt stop us visiting, and the page tells you which two. Blocking us changes nothing about how Google sees your site.
We never need your passwords. There is nothing to install, no plugin, no access to your hosting, and no Google account required. We read the public web, the way anybody else can.

What we cannot see from outside.

This is the part most tools leave out, and it is the reason to trust the rest. A report tells you what is in the way. It cannot tell you what a search engine decided to do afterwards.

Whether Google has actually indexed a page. Having no blocker means a page is allowed to be indexed, not that it has been. Only Google knows, and only through the Google account that manages your site, which you can connect if you want that confirmed.
Whether you appear in any particular AI answer. Nobody outside those companies can see why an answer came out the way it did. What we can tell you is whether their crawlers are allowed to read you at all, which is the part that is actually under your control.
Your Google Business Profile. Kelvix reads your website. It does not read or change your Business Profile, your reviews, or your map listing, even though those matter for a local business too.
Anything behind a login. We see what an anonymous visitor sees. If a page needs a password, it is not a page a search engine is reading either.

Where each finding's authority comes from.

Every check is labelled, and the label is not decoration. It is the difference between a published requirement and an opinion, and you are entitled to know which one you are reading.

19DocumentedSomebody published this as a requirement. The finding carries the link to it.
14MeasuredSomething we read on your own pages, shown to you with the evidence attached.
4Our judgmentNo published rule exists. The finding says so, in those words, on the card.

See all 37 checks

robots.txt and sitemap3

robots.txt declares the same crawler twice
No XML sitemap found
Pages missing from the sitemap

Whether it can be seen at all8

AI assistants are blocked from reading this site
AI training crawlers are blocked
Page points its canonical tag at a different page
Robots rules on this page contradict each other
Page is blocked from being indexed
Page does not return a successful response
Page does not allow text snippets
Pages blocked by robots.txt

Headings and page content4

Page has no main heading
Page has more than one main heading
Headings skip levels
Page has very little text

Links4

Links to pages that no longer exist
Menu links that point nowhere
Links written as http:// on a secure page
Links pass through several redirects

Local business details1

The same phone number is written in different ways

Images3

Images have no alt text
Images have no width or height
Alt text looks auto-generated

Titles and descriptions6

Several pages share the same meta description
Page has no meta description
Several pages share the same title
Page has no title
Page has more than one title tag
Page title repeats itself

Structured data8

Structured data never says what the business is
Structured data does not match what the page shows
Structured data will not parse
Structured data is missing a property Google requires
Page has no structured data
Structured data could include more that Google uses
Structured data uses a type that does not exist
Structured data parses but cannot be used

That is the whole method.

Run it on your own site and see what it says. Free, about a minute, no card.

Check your site

We read your robots.txt first

We read up to 20 pages, one at a time

You get findings, worst first