MeshScore is a 0–100 assessment of recorded site signals related to retrievability, accuracy, and citation authority. It is calculated from the scan's measured checks and explicitly labelled proxies, never an invented industry benchmark. The result is banded from Critical to Excellent and always carries confidence plus unassessed evidence. A lead scan is directional. A deeper scan can still be partial if the target, renderer, or evidence source is unavailable; “full” describes the requested mode, not a guarantee that every signal completed. MeshScore positions and measures site structure, not whether a proprietary engine will cite it.
Think of it as a compact assessment of recorded site signals: one number that summarizes the checks completed for that evidence snapshot, together with confidence and unassessed evidence. It is not a prediction that an engine will cite you. Run a free diagnostic to see yours.
MeshScore summarizes recorded structural signals in one number
Some SEO tools report search-oriented technical, content, link, or visibility signals. MeshScore addresses a narrower question: what retrievability, accuracy, and citation-authority rubric evidence did this EntityMesh scan actually record? It does not answer whether a provider will find, trust, quote, or cite the site.
MeshScore makes the recorded structural evidence repeatable. The calculation is deterministic for the same normalized evidence snapshot. A live rescan can legitimately differ when content, fetch/render coverage, or evidence availability changes, so comparisons must disclose those receipts before attributing movement to the work.
MeshScore is calculated from three components
MeshScore combines three components that sum to 100:
Retrievability (max 33)
Can the bounded scanner retrieve usable public content? This component uses completed crawl, render, and support-hub checks. A client-rendered page may expose little useful HTML to a non-rendering fetcher; the result reports the coverage actually attempted and does not stand in for every external crawler.
Accuracy (max 33)
What consistency and structured-data signals did the scan record? Valid schema can label a page for compatible consumers when the visible content satisfies that type. It does not make a claim safe, correct, selected, or quoted.
Citation Authority (max 34)
Does the page carry the structural signals defined by this rubric? This component uses question-style headings, internal-linking breadth, and the presence of an /llms.txt publisher hint. The component name does not imply that an external engine will use those signals or choose a citation.
Each component is shown on your diagnostic on a /100 scale so they're easy to compare, even though their raw ceilings differ (33/33/34).
The MeshScore bands map the number to a readable status
- 0–29 Critical — Recorded structural coverage is critically limited for the assessed evidence. Foundational review is warranted.
- 30–49 Limited — Some rubric signals are present; material recorded gaps remain.
- 50–69 Developing — The recorded checks show a mixed structural base and reviewable gaps.
- 70–84 Strong — The recorded structural signals are strong for the evidence assessed.
- 85–100 Excellent — labelled "EntityMesh Certified" only when the result is complete, non-proxy, and proof-grade.
Below 70, inspect the completed checks, proxies, and unassessed evidence before choosing work. A diagnostic reports what it measured; it does not promise that every recommendation will change a live rescan or an external-engine outcome. See your band.
You improve MeshScore by fixing retrievability, accuracy, and citation gaps
Improvement maps to the three components:
- Raise Retrievability — make sure key pages render in HTML, fix crawl blocks, and publish a real support/answer hub instead of scattered pages.
- Raise Accuracy — use supported schema types only where the visible page satisfies them, and keep approved answers consistent.
- Raise Citation Authority within this rubric — structure content as questions and answers, link related answers, and publish an
/llms.txtwhere useful. These are measured structural signals, not proof that a proprietary engine will use or cite the content.
EntityMesh can build this infrastructure and run a later comparable scan; movement is shown only from completed evidence rather than assumed.
Compatible before/after scans document recorded movement
MeshScore's calculation is deterministic for the same normalized evidence. A live scan is an observation whose coverage can change even when the operator made no intentional site edit. Each result therefore carries a confidence label and evidence receipt:
- Directional — some inputs are proxies or the evidence coverage is incomplete.
- Proof-grade MeshScore — the complete, non-proxy structural rubric was measured. This grades the assessment evidence, not a ranking, citation, revenue, or causal outcome.
MeshScore assesses recorded structural signals. It does not position a page inside a proprietary provider, predict citation probability, or prove retrieval.
MeshScore complements SEO, AEO, and AI-readiness scores
Your diagnostic also shows SEO Foundation, AEO Readiness, and AI Readiness. These appear beside the related MeshScore component as related signals — they are not added into the composite, because doing so would double-count the same underlying measurements. Items the scan did not attempt are excluded entirely; they never count against you. MeshScore is the unified number; those three are supporting detail. (See also: What is SEO 3.0? and What is GEO?.)
Frequently asked questions
Is a higher MeshScore a guarantee I'll be cited?
No. It assesses recorded site signals. External retrieval and citation must be observed separately, and no tool can guarantee how a provider selects sources.
Why is my score "directional" and not "proof-grade"?
Directional means some inputs are proxies or the evidence coverage is incomplete. It is an evidence boundary, not a claim that a proof-grade structural score has been validated against real AI citations.
Will my score change if I don't change anything?
It can. The scoring function is deterministic, but a live rescan may retrieve different content or coverage because the site, network, renderer, or evidence availability changed. Compare compatible run receipts before attributing a delta.
Ready to measure it? Run your free diagnostic for a confidence-labelled baseline when the bounded scan completes.
Last updated: July 20, 2026.