Skip to content
Signal Layer
Free tool · AI Readability Score

Signal Layer — How well does your page speak to AI?

Score any public page for AI, LLM and generative-search readability in about 15 seconds. We check 46 signals across five categories and show you exactly what to fix, from schema depth and crawler access to answer extractability and citation signals.

Scan mode

Or try an example:

Free while we finalise accounts. Opens the Signal Layer scanner in a new tab.

— what we check

Forty-six signals across five categories.

Each category is weighted by how much it actually moves the needle for AI citation and generative-search surfacing. The full rubric — including every check code and its rationale — is visible inside each scan report.

  1. 01

    Machine Readability

    weight 20 · 12 checks

    Can an AI crawler even reach your content, and is it served as clean, parseable HTML rather than a JavaScript-only shell?

    • JSON-LD structured data present and parseable
    • AI crawler allowlist in robots.txt (GPTBot, ClaudeBot, Perplexity)
    • Content visible without JavaScript (SSR vs CSR)
    • Semantic HTML5 landmarks and language declarations
  2. 02

    Semantic Structure

    weight 25 · 9 checks

    Does the page name the right entity with the right schema, and is the heading hierarchy one an LLM can rebuild from?

    • Primary entity schema matches the page type
    • Single H1 that aligns with user intent
    • Logical H2–H4 hierarchy with no skipped levels
    • BreadcrumbList and FAQPage schema where relevant
  3. 03

    Answer Extractability

    weight 25 · 10 checks

    Can a model lift a direct, self-contained answer out of this page — or does it have to paraphrase across scattered paragraphs?

    • Clear summary paragraph above the fold
    • Key facts upfront, not buried
    • Self-contained 40–120 word answer passages
    • Comparable facts marked up as tables, not prose
  4. 04

    Trust & Provenance

    weight 15 · 7 checks

    Does the content have the citation signals — authorship, dates, organisation schema — that AI systems lean on when deciding what to quote?

    • Author or institutional owner identified
    • Publish and last-updated dates present
    • Publisher Organization schema with sameAs links
    • Content freshness appropriate for the page type
  5. 05

    Retrieval Fitness

    weight 15 · 8 checks

    Is the page discoverable, link-connected, and sufficiently distinct from its neighbours to be worth retrieving in the first place?

    • Page is in sitemap.xml with a lastmod date
    • Descriptive internal anchor text in and out
    • Sufficient unique content density — no thin pages
    • Open Graph + Twitter Card metadata complete

— three scan modes

Pick the right scan for the page.

Not every page needs a full browser render. Most content pages are server-rendered and can be scored in a single HTTP fetch; single-page apps need the headless browser mode. Credits are deducted per scan so you get more mileage by picking the lightest mode that works.

Quick 1 credit
Reads the raw HTML. Fastest. Best for content-first sites where the page is mostly server-rendered.
Smart 1–2 credits
Starts quick; upgrades to a real browser render only if the page looks thin or JavaScript-rendered. The right default for most people.
Full browser 2 credits
Always runs a headless browser render. Use it for React, Vue, Angular or any SPA where the content only exists after JavaScript runs.

— from score to audit

This scores one page. The audit does this for your whole site.

The AI Readability Score runs on a single URL. The Signal Layer AI Content Audit applies the same rubric across your entire content estate — 1,000, 100,000 or 1,000,000+ pages — then prioritises the findings into a roadmap your content, digital and governance teams can action together.

Think of the score as the thin end of the wedge: the free surface to the rigorous analysis. If the free score reveals problems on one page, they'll almost certainly be present across thousands more.

Explore the AI Content Audit

— questions

Answers we give most often.

How is this different from an SEO score?
SEO scores grade for classic search ranking — keyword targeting, internal links, Core Web Vitals. AI readability overlaps at the edges but adds checks that specifically matter to LLMs and generative search: JSON-LD depth, AI crawler allowlists, self-contained answer passages, quotable sentence density, llms.txt manifests, and so on. A page can rank well on Google and still be invisible to ChatGPT, Gemini and Perplexity.
Do you store my URL?
We store the result of the scan so we can return a cached report if the same URL is scored again. We don't store any content from behind authentication and we don't share URLs with third parties. Every result is stamped with the rubric version and prompt version used, so the score remains reproducible.
How accurate is the score?
The rubric combines deterministic heuristics (is robots.txt blocking GPTBot? is there JSON-LD?) with LLM-judged checks (does the H1 match the page's intent? are there self-contained answer passages?). Every check records why it decided what it did, so you can see the reasoning — not just a traffic-light verdict.
What happens when I sign up?
Signing in with Google gives you scan history, shareable report URLs, and more credits than the shared anonymous daily budget. Deeper organisation-wide analysis — across thousands or millions of pages — lives in the Signal Layer AI Content Audit, which applies the same rubric across your entire estate.
Can I embed a score badge on my site?
Not yet. Public report URLs and an embeddable badge are next on the roadmap. For now, the URL in your browser after running a scan can be shared directly.

Score a page now.

Free while we finalise accounts. Fifteen seconds, 46 signals, exact remediation steps — opens the Signal Layer scanner in a new tab.

Want the same rigour across your entire estate?

The AI Content Audit applies this rubric across every page you publish — and turns findings into a prioritised roadmap your teams can action. Book a 20-minute introduction to see what it would look like on your site.

Book an intro call