What We Check
Every check QuickRankAI runs is listed here with its weight and its category. Nothing is hidden behind a paid tier and nothing is scored by a model. If a finding shows up in your result, this is where it came from.
Discovery
Can the page be reached at all. Everything else is moot until the answer is yes.
| Check | Weight | Fails as |
|---|---|---|
| Served over HTTPS | 8 | Structural |
| robots.txt is published | 4 | Quick fix |
| This URL is not disallowed in robots.txt | 10 | Quick fix |
| AI crawlers are allowed | 10 | Quick fix |
| Page is indexable | 10 | Quick fix |
| Canonical URL declared | 5 | Quick fix |
| Canonical points at this site | 5 | Investigate |
| Reached without a redirect chain | 4 | Quick fix |
| XML sitemap is discoverable | 6 | Quick fix |
| Server responded promptly | 5 | Structural |
Page Structure
Whether the document has a readable shape, or is a wall of styled text with no addressable parts.
| Check | Weight | Fails as |
|---|---|---|
| Title tag is present and usable | 8 | Quick fix |
| Meta description is present and usable | 5 | Quick fix |
| Exactly one H1 | 7 | Quick fix |
| Heading levels do not skip | 4 | Quick fix |
| Page is broken into sections | 5 | Quick fix |
| Page links onward into the site | 6 | Quick fix, or structural at zero links |
| Anchor text describes the destination | 4 | Quick fix |
| Images carry alt text | 4 | Quick fix |
| Mobile viewport declared | 3 | Quick fix |
Entity Clarity
Whether the page states who published it and what it is, in a form a machine reads rather than infers.
| Check | Weight | Fails as |
|---|---|---|
| Structured data is present | 10 | Quick fix |
| Publisher is declared as an entity | 9 | Quick fix |
| Entity is linked to outside references | 7 | Quick fix |
| Page type is declared | 6 | Quick fix |
| Structured data uses stable identifiers | 5 | Quick fix |
| Naming is consistent across signals | 4 | Investigate |
AI Readiness
Whether the content is actually in the response, and whether anything can separate it from the furniture around it.
| Check | Weight | Fails as |
|---|---|---|
| Content is in the HTML, not assembled by script | 12 | Structural |
| Page has enough readable text to be useful | 8 | Quick fix, or structural under 80 words |
| Main content is inside a semantic landmark | 6 | Quick fix |
| Body text survives boilerplate stripping | 7 | Quick fix, or structural at very low ratios |
| Document language is declared | 3 | Quick fix |
| Open Graph metadata present | 3 | Quick fix |
| llms.txt published | 4 | Quick fix |
Content Signals
Whether the page answers something, and whether it does so in a shape anything can reuse.
| Check | Weight | Fails as |
|---|---|---|
| Opens with a real answer | 9 | Quick fix |
| Claims are specific rather than general | 6 | Quick fix |
| Question and answer structure is marked up | 5 | Quick fix |
| Page carries a freshness signal | 5 | Quick fix |
| Page cites something outside itself | 4 | Quick fix |
| Headings describe their sections | 4 | Quick fix |
What a quick inspection cannot tell you
This section exists because the tool would be dishonest without it.
QuickRankAI reads one public URL and the three or four small files sitting beside it. That is the entire evidence base. From that it can identify a great many common technical, structural, entity, content, and machine-readable problems, and it can tell them apart from each other. Here is what it cannot do:
- It cannot see your site. One URL is one URL. The other four thousand pages built from the same template might be fine or might be worse, and nothing in this result says which.
- It cannot see your history. No Search Console data, no analytics, no server logs, no record of what crawlers actually did on your site last month.
- It cannot see competition. A page can score 96 and never be found, because forty better-established pages already answer the same question better.
- It cannot see demand. Nothing here knows whether anyone is looking for your subject.
- It cannot execute your JavaScript. That is deliberate, because most AI retrievers do not either, but it means a page that renders fine in a browser can inspect badly, and the inspection is right about that.
- It cannot promise anything. Not rankings, not AI citations, not traffic. Fixing every finding makes a page more readable and more reachable. What happens after that depends on things no page-level tool can observe.
A quick inspection catches the obvious problems, and the obvious problems are worth catching. It is not proof that the whole system is healthy. Replacing the filter helps. It does not rebuild the engine.