Skip to content
(770) 742-9385
100% U.S.-Based TeamIn God We Trust
Nordax DigitalFree Visibility Score
Nordax Digital

NDVAS · Open specification

Visibility starts
with structure.

The Nordax Digital Visibility Architecture Specification. A published standard for building websites that people, search engines, and AI systems can understand.

Visibility architecturev1.4.1

Level 1 · Public specification

Foundational Visibility Architecture

The technical foundation: clear structure, accessible content, consistent identity, and machine-readable business information.

  • Security & metadata
  • Structured data & crawlability
  • Performance & semantic content
One connected framework.Three levels.
Current version
1.4.1 / Technical revision
Level 1 access
Public Distribution Approved
Updated
Created by
Alexz Shepherd / Nordax Digital

A shared standard

Clarity in the code.
Confidence in the foundation.

The Nordax Digital Visibility Architecture Specification defines the required technical, structural, and semantic implementation standards for all websites deployed by Nordax Digital.

It connects traditional search foundations with machine-readable identity and content, so the website and the business behind it tell a consistent story.

How to read the specification

Three requirement tiers.

REQUIRED

The implementation baseline.

Nordax acceptance requirements. Record applicability and any reviewed, approved exception.

RECOMMENDED

Useful when applicable.

Adopt when supported by the project’s needs, implementation context, and evidence.

EXPERIMENTAL

Evaluate for a real use.

Test the consuming system and record limitations before treating an emerging convention as a dependency.

Download the public Level 1 specification (PDF)

Technical revision 1.4.1What changed

Based on the expanded v1.4 standard from June 2026. Reviewed and corrected October 19, 2026.

  • Expanded the earlier website’s 12 implementation areas to the 20 areas in v1.4.
  • Distinguished Nordax’s 2.0-second LCP target from Google’s published 2.5-second good threshold.
  • Separated search discovery, AI retrieval, and owner-controlled training permissions.
  • Corrected FAQ / review guidance, sameAs identity links, and optional machine-facing files.
  • Removed unsupported ranking guarantees, invented performance metrics, and fixed content-length formulas.
  • Kept Level 2 and Level 3 implementation details internal and added primary-source references.

Level 1 · Version 1.4.1

The foundation.
In full.

Explore all 20 public implementation areas, including local, international, accessibility, content, and deployment requirements. Select a section to go straight to the details.

Jump to a section

Section 01

Security & trust signals

Protect visitors and make the website a dependable destination.

REQUIRED
  • Enforce HTTPS and redirect alternate hostnames to the chosen canonical hostname.
  • Set X-Content-Type-Options: nosniff and an explicit Referrer-Policy.
  • Apply an appropriate Content-Security-Policy, using nonces or hashes where needed; test third-party integrations before enforcement.
  • Restrict framing through CSP frame-ancestors and, where compatible, X-Frame-Options. Document any legitimate embedding requirements.
  • Configure Permissions-Policy for camera, microphone, geolocation, and other capabilities according to actual product needs.
  • Enable HSTS after validating HTTPS coverage. Use includeSubDomains and preload only when every relevant subdomain can support the commitment.
  • Remove avoidable technology disclosure headers. Monitor for compromised content, deceptive redirects, and broken security controls.
  • Keep essential content accessible without intrusive interstitials.
RECOMMENDED
  • Review DNS prefetch and external connection hints against the real performance profile.

Section 02

Core metadata architecture

Give each indexable page a clear identity and accurate previews.

REQUIRED
  • Define the production metadata base URL and a consistent title template.
  • Provide a descriptive, unique title, meta description, and canonical URL for each indexable page.
  • Identify the business, application, publisher, and language accurately.
  • Provide appropriate Open Graph and social-card metadata with real, accessible images and descriptive alternative text.
  • Keep publication and modification dates truthful and consistent with visible content.
RECOMMENDED
  • Allow large image previews where appropriate with max-image-preview:large. Use 1200-pixel-wide editorial preview images when they suit the placement.
  • Validate metadata in rendered HTML and production social previews.

Section 03

Structured data implementation

Describe real businesses and content with accurate, connected records.

REQUIRED
  • Publish Organization and WebSite records with stable identifiers, the canonical URL, and supported properties such as name, legalName, inLanguage, publisher, and contactPoint.
  • Use LocalBusiness only where applicable, with truthful location, service-area, contact, and opening information.
  • Use Article or a suitable subtype on editorial content, with genuine authorship, publication dates, images, and publisher references.
  • Provide BreadcrumbList for meaningful page hierarchies. Validate syntax and supported property names at validator.schema.org.
  • Ensure markup agrees with content people can see. Test Google-supported rich-result types against the relevant feature guidelines.
RECOMMENDED
  • Use Service, Person, ImageObject, and meaningful isPartOf / hasPart relationships where appropriate.
  • Use FAQPage for genuine visible questions and answers when useful to a consuming system.
  • Use Review or AggregateRating only for an appropriate entity and compliant, genuine, visible review data.
EXPERIMENTAL
  • Evaluate Product, VideoObject, Event, DefinedTermSet, HowTo, or Speakable for a real content need and a verified consuming system.

References [7][6]

Section 04

Search & AI crawler policies

Make deliberate decisions about search discovery, retrieval, and training.

REQUIRED
  • Keep robots.txt reachable and syntactically valid. Include the sitemap location and the intended crawl rules.
  • Document separate owner-approved policies for search indexing, AI retrieval, model training, and user-initiated access.
  • Allow the search crawlers needed for the agreed discovery strategy, and check CDN / WAF rules against those decisions.
  • Verify current provider documentation, agent names, and supported IP verification before creating allow rules. Do not treat a user-agent string as authentication.
RECOMMENDED
  • Review Googlebot for Google Search; OAI-SearchBot for ChatGPT search; and provider-documented retrieval crawlers as relevant to the project.
  • Record the owner’s independent decisions about training-related controls such as GPTBot and Google-Extended.
EXPERIMENTAL
  • Maintain llms.txt, llms-full.txt, or an agent-facing catalog for identified consumers, with ownership and update tests.
Examples of separate provider controls
ControlPurposePolicy
GooglebotGoogle Search crawlingAgreed public indexing strategy
OAI-SearchBotChatGPT searchAgreed search discovery strategy
GPTBotPotential model-training useIndependent owner decision
Google-ExtendedOther Google AI usesSeparate from Search inclusion
ChatGPT-UserUser-initiated accessrobots.txt rules may not apply

References [3][4][5]

Section 05

Crawlability, indexation & crawl health

Keep public pages discoverable and private systems protected.

REQUIRED
  • Generate an up-to-date XML sitemap containing canonical URLs intended for indexing.
  • Link important pages through ordinary crawlable navigation and relevant contextual links.
  • Use appropriate canonicals, redirects, status codes, and parameter handling. Return a real 404 or 410 when removed content has no relevant replacement.
  • Protect staging, administrative, client, and confidential areas with access controls. Use noindex where appropriate for non-private pages excluded from search.
  • Monitor meaningful 4xx / 5xx errors and avoid unnecessary redirect chains and loops.
RECOMMENDED
  • Use image or video sitemap extensions where relevant. Configure X-Robots-Tag for non-HTML responses when needed.
  • Review crawl logs and Search Console in proportion to the size and change rate of the website.
EXPERIMENTAL
  • Evaluate IndexNow for a supported search engine and deployment workflow.

References [10][5]

Section 06

Performance & Core Web Vitals

Set measurable performance budgets and preserve usable content.

REQUIRED
  • Use LCP under 2.0 seconds as the Nordax project target. Assess INP against 200 ms and CLS against 0.1, using representative devices and field data where available.
  • Render essential business and editorial content in server-delivered HTML or an equivalently accessible output; do not make discovery depend on an untested rendering capability.
  • Serve properly sized images, reserve their dimensions, and avoid lazy-loading the principal above-the-fold image.
  • Control JavaScript and font cost, avoid hydration shifts, and configure caching with a tested invalidation strategy.
  • Check mobile layout, keyboard interaction, and essential tasks under realistic network conditions.
RECOMMENDED
  • Measure field performance at the 75th percentile where sufficient data exists. Retain lab diagnostics for debugging and prelaunch checks.
  • Preload only resources that are truly critical; review third-party scripts and embedded media.

References [8]

Section 07

Image & visual search optimization

Make real imagery useful, accessible, and efficient.

REQUIRED
  • Write descriptive alt text for informative images and empty alt text for decorative images.
  • Use meaningful filenames where practical, reserve image dimensions or aspect ratios, and preserve the intended composition.
  • Serve suitable dimensions and compression, with formats and fallbacks supported by the delivery platform.
  • Load below-fold imagery lazily and prioritize the image responsible for the opening experience.
RECOMMENDED
  • Connect significant imagery to relevant page copy. Add accurate image metadata and sitemap entries where appropriate.
  • Use large, high-quality editorial previews where a distribution feature benefits from them.

Section 08

Content structure, semantics & demonstrated expertise

Make content clear to people and to machines that process it.

REQUIRED
  • Use a clear primary heading, logical heading structure, and appropriate main, article, section, aside, and nav elements.
  • Use descriptive link text and relevant internal links.
  • Show genuine authorship and accurate dates for editorial material. Attribute claims, quotes, and expertise to the right people.
  • Keep service descriptions, business identity, and contact information clear and consistent.
RECOMMENDED
  • Use original examples, documented processes, photography, and first-hand evidence where they help the reader.
  • Show meaningful update or revision information and answer real customer questions in plain language.

Section 09

Content clusters & topical architecture

Connect relevant material so readers can explore a subject.

REQUIRED
  • Organize core service or topic areas into useful overview pages with contextual links to supporting material.
  • Link supporting pages back to the relevant overview and to related pages when useful.
  • Identify and resolve important orphaned pages without manufacturing duplicate or thin content.
RECOMMENDED
  • Set content depth and publishing cadence from the audience’s needs, available expertise, and real business priorities.
  • Review internal links and navigation depth as the library grows.

Section 10

Answer clarity & generative search readiness

Make supported answers easy to find and understand.

REQUIRED
  • Answer the main question directly and supply enough context for the answer to remain accurate when quoted.
  • Attribute data and material claims to reliable sources. Distinguish facts, estimates, opinions, and owner-reported results.
  • Keep service, location, availability, and identity details consistent with the underlying business.
RECOMMENDED
  • Use concise explanations, useful comparisons, steps, tables, and original evidence where they suit the subject.
  • Evaluate answer visibility with a documented query set and repeatable observations.

References [1][2]

Section 11

Canonical authority control

Maintain a coherent public URL structure.

REQUIRED
  • Enforce the intended HTTPS hostname and a consistent policy for duplicate URL variants.
  • Provide accurate canonical references for indexable pages; keep internal links and sitemap URLs aligned.
  • Map changed URLs to genuinely equivalent destinations. Avoid redirecting unrelated retired content to the homepage.
  • Give distinct paginated pages appropriate canonical treatment rather than collapsing every page onto page one.
RECOMMENDED
  • Test casing, trailing slashes, tracking parameters, redirect loops, and representative paginated or filtered routes.

Section 12

Knowledge graph & entity signals

Describe the business as a consistent, identifiable entity.

REQUIRED
  • Maintain stable identifiers for the organization and relevant services, people, and real locations.
  • Keep identity, contact, and service facts consistent across the website and maintained external profiles.
  • Use sameAs only for pages that unambiguously identify the same entity; use other suitable links for coverage or references.
RECOMMENDED
  • Maintain an accurate entity home and appropriate business profiles.
  • Review Knowledge Panel information if a panel exists. Evaluate Wikidata participation only when its notability criteria are met.

References [9][13]

Section 13

Local & multi-location architecture

Represent genuine places and service areas accurately.

Applies to businesses with an eligible physical location or local service area.

REQUIRED
  • Provide useful, distinct information for real locations and service areas. Avoid copy-and-paste city pages created only to capture searches.
  • Use truthful LocalBusiness details and business-profile eligibility. Do not present a mailing address or virtual office as a staffed, customer-facing location.
  • Maintain accurate contact details, services, hours, and location references across relevant owned profiles.
RECOMMENDED
  • Build relevant local and industry profiles rather than pursuing a directory-count quota.
  • Request genuine feedback in line with each platform’s policies and keep a practical response process.
  • Add useful location-specific service details, team information, FAQs, and genuine proof.

Section 14

International SEO

Connect equivalent language and regional experiences.

Applies to projects targeting more than one language or country.

REQUIRED
  • Use valid, reciprocal hreflang references, including each page itself, for genuinely equivalent language or regional alternatives.
  • Validate alternate URLs, canonical relationships, language / region codes, and the selected HTML, HTTP-header, or sitemap method.
  • Keep translated content and business facts accurate, with appropriate regional variation.
RECOMMENDED
  • Choose URL structure from business and operational needs. Use x-default where it describes a real fallback.
  • Automate alternate-link checks as the number of locales grows; maintain locale-aware identity records.

References [11]

Section 15

Accessibility & inclusive design

Make essential content and interactions usable across access needs.

REQUIRED
  • Provide meaningful text alternatives, native controls where suitable, clear labels, and a logical reading and focus order.
  • Support keyboard navigation, visible focus, understandable form errors, and accessible dialogs or menus.
  • Keep content readable when zoomed or reflowed, respect reduced-motion preferences, and avoid color-only status cues.
  • Audit new templates against the agreed accessibility target with automated checks and manual keyboard / assistive-technology review.
RECOMMENDED
  • Use WCAG 2.2 Level AA as the target for new work and record any exceptions or remaining issues.
  • Use at least 4.5:1 contrast for normal text and 3:1 for large text, with applicable non-text contrast checks.
  • Prefer comfortable 44-48 CSS-pixel touch targets where practical. Apply the relevant WCAG minimum size / spacing rules and exceptions.

References [12]

Section 16

Digital PR & earned authority

Earn relevant attention through useful work and credible evidence.

REQUIRED
  • Avoid manipulative link schemes, fabricated coverage, and undisclosed paid endorsements.
  • Attribute research and expert commentary accurately. Distinguish paid placements from editorial decisions.
  • Review material link or reputation issues using an evidence-based remediation process.
RECOMMENDED
  • Create useful original research, documented case studies, and genuinely helpful expert contributions.
  • Track relevant coverage, referral traffic, and business outcomes. Assess editorial quality and audience relevance.

Section 17

AI interpretation & ongoing observation

Maintain reliable information and evaluate actual discovery behavior.

REQUIRED
  • Publish accurate, current descriptions of services and the business in the formats the project supports.
  • Keep authorship, source attribution, dates, and important entity references consistent with visible content.
  • Document which public endpoints, optional files, integrations, and monitoring surfaces are actually implemented.
RECOMMENDED
  • Track brand mentions and linked citations separately, using a recorded query set, platform / mode, locale, and observation date.
  • Use observed gaps to improve useful content and identity accuracy; distinguish correlation from causation.

References [2]

Section 18

NIST alignment

Make any framework mapping explicit and reviewable.

REQUIRED
  • Where NIST alignment is claimed, maintain a written mapping from relevant NDVAS controls to the cited AI RMF guidance, with scope, evidence, and limitations.
  • Keep implementation evidence and the responsible owner identifiable.

References [14]

Section 19

Deployment validation

Confirm the agreed implementation with reproducible evidence.

REQUIRED
  • Validate metadata, supported structured data, canonicals, sitemap output, intended crawler policies, HTTPS, and material redirect behavior.
  • Check representative page templates and real user journeys on desktop and mobile, including loading, error, and confirmation states.
  • Measure the agreed performance budgets and complete accessibility checks for the tested scope.
  • Validate locale and real-location requirements where applicable. Check that private material cannot be reached through public routes.
  • Record results, release owner, remaining issues, and reviewed exceptions. REQUIRED items must pass or have a documented, approved exception before claiming compliance.
RECOMMENDED
  • Retain a dated evidence pack and configure monitoring for meaningful failures after release.

Section 20

Nordax Digital compliance standard

State what conformance covers and who maintains it.

REQUIRED
  • Apply the relevant Level 1 controls to Nordax Digital website work. Record non-applicable controls and approved project-specific exceptions.
  • Identify the specification version, implementation scope, acceptance evidence, and owner for each compliance claim.
  • Review the implementation after material platform, content, or business changes.

Appendix A1

NDVAS Levels 2 & 3

The progression beyond the public foundation.

Level 1 is the public specification. The detailed Level 2 and Level 3 implementation standards remain internal. These summaries describe their scope.

Level 1

Foundational Visibility Architecture

The technical foundation: clear structure, accessible content, consistent identity, and machine-readable business information.

Level 2

Autonomous Visibility Architecture

Systems that maintain the foundation as your website evolves, keeping visibility signals current when content changes.

Level 3

Autonomous Authority Infrastructure

Maintained entity records, verified identity information, supported integrations, and observation of how the business is discovered.

Level 2 maintains the implemented visibility signals as content changes. Level 3 connects maintained entity information, verification, supported integrations, and observation. For Nordax-managed Level 3 deployments, the platform integration is delivered through nordax.ai.

Implementation scope and evidence determine conformance. Recognition, rankings, and citations are decisions made by external systems.

Appendix A2

Questions & answers

A clearer view of the standard and its scope.

What is NDVAS?

NDVAS is the Nordax Digital Visibility Architecture Specification: our implementation standard for website structure, performance, accessibility, business identity, and machine-readable information.

Who created NDVAS?

NDVAS was created by Alexz Shepherd at Nordax Digital. Level 1 is published as a reference for the websites we build and maintain.

What are the three levels?

Level 1 is Foundational Visibility Architecture. Level 2 is Autonomous Visibility Architecture, covering ongoing maintenance. Level 3 is Autonomous Authority Infrastructure, covering managed entity infrastructure, integrations, and observation.

Which parts are public?

The Level 1 specification is public. The detailed Level 2 and Level 3 implementation standards remain internal. Public summaries explain their scope without publishing their confidential requirements.

Does NDVAS guarantee rankings or AI citations?

No. NDVAS defines implementation and validation requirements. Search engines and AI systems decide what to index, surface, or cite; meeting a technical standard does not guarantee those outcomes.

Is NDVAS endorsed by NIST?

No. NIST has not validated, certified, or endorsed NDVAS. Any NIST alignment claim must be supported by a scoped, documented mapping to the relevant guidance.

Appendix A3

Sources & references

Primary documentation reviewed October 19, 2026. Provider guidance may change.

  1. [1]Google Search Central: AI features and your website
  2. [2]Google Search Central: Generative AI optimization guidance
  3. [3]OpenAI: Overview of crawlers
  4. [4]Google: Common crawlers and Google-Extended
  5. [5]Google: robots.txt specification
  6. [6]Google: Review snippet structured data guidelines
  7. [7]Google Search Central: Documentation updates, including FAQ retirement
  8. [8]Google: Core Web Vitals
  9. [9]Schema.org: sameAs identity references
  10. [10]Google: Build and submit a sitemap
  11. [11]Google: Localized versions of your pages
  12. [12]W3C: Web Content Accessibility Guidelines 2.2
  13. [13]Wikidata: Notability policy
  14. [14]NIST: AI Risk Management Framework

Public Level 1 · Technical revision

NDVAS 1.4.1. Based on v1.4 from June 2026, with technical corrections and primary references reviewed October 19, 2026.

Nordax Digital · Alexz Shepherd · Canonical: nordaxdigital.com/ndvas

Put the standard to work

A stronger foundation.
For your business.

NDVAS is the standard we build to. Start with an assessment to see where your website stands and what to improve next.