SEO
Your site’s SEO is set by the system. No redesign and no AI can take it away.
Title, description, canonical URL, Open Graph and structured data come from a part of the code no theme touches. Change the whole design, by hand or by typing, and Google keeps reading the same thing. The rest — analysis, sitemaps, redirects, 404s — works on its own and costs no credits.
What you get
What’s a paid plugin elsewhere is always on here
A site served without a title or a canonical URL is broken, not “missing a module”. That’s why SEO comes with every plan, the free one included, and everything worked out by rules never spends a single credit.
-
A scored analysis, no AI and no cost
On save it tells you the title runs long, there are two H1s and three images are missing alt text. Thirty rule-based checks: the same page gets the same score today and a month from now.
-
Sitemaps with real dates
The core generates them, because it’s what knows what’s published: by type, paginated and with images. The lastmod is the day you changed the page, not today, which is what makes half the sitemaps out there useless.
-
Redirects that create themselves: Every change is a version
Swap “/services-2024” for “/services” and the old address answers 301, and so do the pages that hung from it. Anything pointing to the old one moves to the new one without chaining hops, and a rule you wrote by hand is never overwritten.
-
The 404s that matter
A site out in the open gets hit all day with requests for /wp-login.php and /.env. That noise is filtered out, so the list you read is the addresses real people are looking for, with a button that turns each one into a redirect.
-
Orphan pages and broken links
On publish, what links to what is recalculated. That’s where you get the list of pages not a single link of yours reaches — the ones Google barely visits — and the list of links pointing at something that no longer exists.
-
llms.txt and AI crawlers
An llms.txt generated from your published content, for the assistants that answer instead of Google today. And a switch per crawler (GPTBot, ClaudeBot, PerplexityBot…) with the consequence of each block written next to it.
The head
The title, the canonical and the structured data are drawn by the system. The theme never sees them.
The theme hands over the whole document and the system decides what goes in the head. That code lives in a folder no theme can touch: not writing it yourself, not us writing it, and not asking the AI with the best of intentions.
-
Out of reach of any prompt
“Clean up the header code” sounds reasonable and pulls a site out of Google. That folder is blacklisted, in the code, and it’s checked twice.
-
Structured data writes itself
Article, organization, breadcrumb, product with its price and availability. And if a page has an FAQ block, its FAQPage comes from there with nothing to configure.
-
You never share a bare link
No social image set? The document’s cover is used; no cover either, and it falls back to the one in your brand settings. On WhatsApp and LinkedIn something always shows.
<title>Zapatillas Runner · Tienda San José</title><link rel="canonical" href="https://clinicasanjose.pe/tienda/zapatillas-runner"><meta property="og:image" content="…/zapatillas-runner-og.jpg"><script type="application/ld+json">{"@type":"Product","offers":{"price":"189.00","priceCurrency":"PEN"}}</script>How it works
From publishing a page to Google having it
Nothing to install, nothing to configure and no credits spent in any of the four steps.
-
You write and see the score
The checks run with the save itself, so they never add a separate version to your history. Each one says what to change, not what’s wrong.
SEO 82 / 100
-
You publish
The system writes the head from your data, puts the page in the sitemap with its real date and recalculates what links to what across your site.
lastmod: the day it changed
-
The search engines are told
IndexNow pings Bing and compatible engines the moment you publish. Google isn’t pushed anything: it’s left to read the sitemap, served on your own domain.
sitemap.xml · robots.txt · llms.txt
-
You change the URL and nothing breaks
Change the slug of a published page tomorrow and the old one answers 301 to the new one, and so do the pages that hung from it. Nobody hits a 404 because you reorganized.
301 · created on its own
What you mark as noindex stays out of the sitemap by definition, not by configuration: there’s no switch that puts it back in. And your store’s cart stays out too, without you having to remember.
Redesign as often as you like. The title, the canonical and the structured data don’t move.
The proof
This is what an AI redesign changes. And this is what never shows up in the diff.
When you ask for “a more serious look across the whole site”, what gets rewritten is tokens, styles and layouts. The head isn’t in the diff because the theme doesn’t write it: it hands the document to the system, and the system sets the title, the canonical URL and the JSON-LD. There’s no way to prompt them away.
- Article, organization, breadcrumb and product with offer, filled in from the document
- A page with an FAQ block generates its FAQPage on its own, with nothing to configure
- Structured data is validated on save; the theme never writes JSON-LD by hand
- The score travels with the save: your history keeps your changes, not a number’s
change «the button in yellow»
change «a dark background»
change «the title on the left»
change «a bigger title»
change «a rounded button»
The deterministic part
What runs on every save, without calling a single model
- checks: keyword in the title, URL, H1 and first paragraph; lengths; a single H1; alt text; links; readability; duplicate titles
- 30 checks: keyword in the title, URL, H1 and first paragraph; lengths; a single H1; alt text; links; readability; duplicate titles no randomness, no dates, no network
- as a score, with a traffic light and the list of what’s missing. What couldn’t be measured doesn’t count against you, because penalizing something nobody looked at is lying with a number
- 0–100 as a score, with a traffic light and the list of what’s missing. What couldn’t be measured doesn’t count against you, because penalizing something nobody looked at is lying with a number the same today and a month from now
- credits for all of the above. With no balance left, the analysis, the score, the sitemaps and the redirects keep working exactly the same
- 0 credits for all of the above. With no balance left, the analysis, the score, the sitemaps and the redirects keep working exactly the same only suggestions cost anything
- generated from your published content, alongside the sitemap with real dates and a robots.txt that warns you when you get it wrong
- llms.txt generated from your published content, alongside the sitemap with real dates and a robots.txt that warns you when you get it wrong for the ones that answer instead of searching
SEO assistant
The AI proposes titles, descriptions and alt text. Accepting is still yours.
Several title and description options from the real content, a focus keyword, alt text for images in bulk — which clears a whole site’s backlog in an afternoon — FAQs, and links pointing at your orphan pages. Every suggestion is accepted, edited or discarded, one at a time. It doesn’t invent awards or figures that aren’t in the document.
- It spends the same credits as design, with its prompt and its model in the audit trail
- In bulk the cost is estimated and confirmed before anything is spent: a hundred descriptions cost money
- The score never depends on the model: the AI explains each failure in plain English, it doesn’t grade
- If there’s nothing to suggest from, it says so instead of inventing it, and doesn’t charge
Balance
2,450 credits
Used this cycle 68%
Recent changes
-
“main button in yellow”
this section · 1 credit
-
“form at the top”
this page · 3 credits
-
“redesign rejected by policy”
0 credits (refunded)
Questions
What people ask about SEO
Does Google penalize a site built with AI?
What Google reads is the title, the description, the canonical URL, the structured data and the content, and the system sets all of that from your data, not the model. The AI changes how your site is drawn, never what it says or how it describes itself to the search engine. A redesign by typing and a redesign by hand leave the same head.
Do I need a separate SEO plugin?
No. Templated title, description, canonical URL, Open Graph, structured data, sitemaps, robots.txt, redirects and the scored analysis are built in, on every plan. What’s worked out by rules is free; only the assistant’s suggestions spend credits.
Can I write the title and description for each page myself?
Yes, and that’s the normal way: every page, post and product has its own SEO fields, with a preview of how it will look on Google as you type. Leave the title empty and it’s built from the template in your settings (“{title} · Your brand”), so no page ever goes out without one.
How is the SEO score calculated?
With thirty rule-based checks, no model: keyword, lengths, a single H1 and a correct heading hierarchy, images without alt text, internal and external links, readability, duplicate titles, canonical URL and orphan pages. What couldn’t be measured doesn’t count against you. The same page gets the same score today and a month from now, and every check says what to change, not what’s wrong.
What happens if I change a page’s URL?
If it was published, the old URL answers 301 to the new one, and so do the pages that hung from it. Rules that pointed to the old address are updated so hops don’t chain. You can turn it off per page, and a redirect of yours is never overwritten.
Can I block ChatGPT and the rest?
Yes, with a switch per crawler (GPTBot, ClaudeBot, PerplexityBot, CCBot) in your robots.txt. They’re allowed by default: for almost all our customers, being cited in an answer is traffic, and blocking them doesn’t change your position on Google. The consequence is written next to each switch.
Does the assistant write my copy and publish it?
It proposes, and applying is a separate action you take with your own permissions. It works with what’s in the document: it doesn’t add awards, certifications or figures nobody wrote. When you ask for a hundred descriptions, it tells you what it will cost before starting, and if there’s nothing to suggest from, it says so and doesn’t charge.
What if I edit robots.txt and get it wrong?
You can edit it; it’s your site. But you get warnings as you type: a stray Disallow: / pulls an entire site out of Google and nobody notices until visitors stop coming. Every change to robots.txt and to the crawler policy is logged in the audit trail.
Redesign as many times as you like. Your rankings stay where they are.
The head, the sitemaps and the redirects are set by the system, on every plan. Start free, no credit card.