# AiVIS Cite Ledger, See If AI Actually Cites Your Site AiVIS Cite Ledger is an evidence-backed AI citation readiness audit. It measures how ChatGPT, Perplexity, Claude, and Google AI interpret, trust, and cite a website, then returns a 0–100 CITE LEDGER score with BRAG evidence-linked findings and prioritized fixes. The platform is not a traffic dashboard. It is a deterministic evidence pipeline that captures crawlable page signals, evaluates answer-engine citation behavior, commits citation records to an immutable ledger, and derives registry scores from committed evidence. ## Gap: AI answer engines need extractable citation evidence ### Traditional search rank does not prove answer-engine inclusion A website can rank in traditional search while still failing AI citation readiness. Answer engines need clear entity definitions, structured data, crawlable text, canonical URLs, source trust signals, and answer-ready sections that make attribution unambiguous. ### Unstructured pages create citation gaps Common blockers include missing Organization or WebPage schema, weak FAQ coverage, vague product definitions, thin body copy, duplicate or absent canonical tags, broken heading hierarchy, inaccessible robots or sitemap signals, and absent external identity references. ### Uncited claims are treated as evidence gaps AiVIS Cite Ledger labels unsupported AI claims as uncited instead of converting them into fabricated authority claims. The system downgrades uncited output, exposes the gap in the report, and ties the recommended fix to the specific missing evidence. ## Evidence: the CITE LEDGER records what was found ### Each audit starts with live extraction The scan captures HTML, metadata, JSON-LD schema, headings, internal links, canonical tags, content depth, policy signals, and crawlability evidence from the submitted URL. Extracted evidence becomes the source material for scoring and remediation. ### Parallel signal checks test citation behavior The signal layer evaluates AI-model interpretation, web-search corroboration, instant-answer signals, mention sources, and SERP evidence where the workspace tier allows it. These checks identify whether the entity is present, absent, displaced, or cited without enough support. ### Ledger rows are the source of truth Committed CITE LEDGER records preserve the query, model or source context, URL evidence, citation state, and hash-locked traceability root. Registry metrics such as visibility score, authority score, entity clarity, and query coverage are derived from those committed rows. ## Fix: recommendations are ranked by citation impact ### Schema fixes improve machine-readable identity AiVIS prioritizes Organization, WebSite, WebPage, SoftwareApplication, FAQPage, DefinedTerm, BreadcrumbList, and product or service schema when those structures match the page intent. The goal is entity clarity, not schema stuffing. ### Content fixes make answers easier to extract Recommended content changes focus on answer-first definitions, short evidence-backed explanations, gap-evidence-fix sections, product boundaries, tier language, methodology summaries, and FAQ answers that align with the visible page and JSON-LD. ### Technical fixes protect attribution Canonical URLs, HTTPS, semantic landmarks, image alt text, internal links, robots policy, llms.txt guidance, sitemap freshness, and policy or contact pages help crawlers verify the page and connect the entity to the correct publisher. ## Ten weighted CITE LEDGER scoring families (avs-v3) ### Schema & Structured Data — 18% JSON-LD presence and completeness across all schema types. Checks entity references, relationship coverage, syntax validity, and multi-type diversity. Missing JSON-LD is a hard blocker that caps the composite score at 79. ### Entity & Heading Signals — 14% Whether the page declares a clear, unambiguous entity identity. Checks single H1 presence (hard blocker if absent), H2 section structure, title-to-OG title semantic consistency, and entity declaration clarity. ### Authority & E-E-A-T — 12% Off-page authority signals that AI systems use to evaluate trustworthiness. Checks Knowledge Graph entity record, SERP top-10 organic presence, featured snippet or knowledge panel ownership, and external source citations on-page. ### Meta Tags & Open Graph — 10% Metadata completeness for AI extraction and social graph parsing. Checks title tag (hard blocker if absent), meta description, OG title, OG description, OG image, and HTML lang attribute. ### Content Depth — 10% Whether the page contains enough substance for AI systems to extract and cite. Checks word count of at least 300, question-style H2 headings for snippet eligibility, and TL;DR or summary block presence. ### Crawlability & Bot Access — 8% Whether AI crawlers can access and retrieve this page. Checks robots.txt policy, AI crawler permissions (hard blocker if blocked), and llms.txt advisory file presence. ### Renderability & Page Speed — 8% Technical rendering signals that affect AI retrieval reliability. Checks page load under 3 seconds, LCP under 2500 ms, image presence, and image alt text coverage of at least 80%. ### Citation Signal Quality — 8% Signals that make the page easier for AI systems to quote and attribute. Checks meta description length (25–160 chars), meta/OG description consistency, and TL;DR positioned near the top of the page. ### Indexability & Link Graph — 6% Link graph signals that reinforce page authority and discoverability. Checks canonical URL presence, internal links, external links, and XML sitemap accessibility. ### Security & Trust Signals — 6% Trust infrastructure that AI systems verify before citing a source. Checks HTTPS canonical URL, language and hreflang targeting, and absence of evidence contradictions. ## Methodology summary ### Input validation and extraction Every analysis begins with a safe URL, sanitized user-supplied fields, and schema validation before the page is fetched and parsed. The extraction step builds the entity graph used by later checks. ### Citation resolution gate AI claims must be matched to real page evidence, web results, or SERP signals. If support is missing, the claim is labeled uncited and surfaced as a gap rather than treated as verified authority. ### Immutable ledger and registry derivation The ledger commit creates stable scan traceability. Registry values are read-only aggregates computed from ledger rows, which keeps public reports reproducible and prevents client-authored score overrides. ## Actor boundaries and audit scale ### Platform owner controls stay internal The platform owner operates infrastructure, policy, orchestration, billing integration, and system-level controls. Those controls are not described as workspace-member capabilities because customer-facing pages must distinguish operator responsibilities from workspace features. ### Workspace members receive evidence views and actions A workspace member submits URLs, reviews citation readiness reports, exports evidence, compares gaps, and applies remediation guidance. The member sees the scan result, live execution state, and evidence-backed output rather than an invented persistent dashboard metric. ### Agent runtime performs background execution Automated workers, schedulers, queues, and browser execution loops perform extraction, probing, visibility observation, drift checks, and publication tasks. Runtime internals are surfaced only as trace metadata where that information helps verify the scan. ### Large audits separate discovery, sampling, and deep analysis For large websites, AiVIS separates broad URL discovery from deterministic sampling and expensive deep analysis. Public telemetry should distinguish URLs discovered, pages sampled, pages fully audited, sampling strategy, and coverage so a cap is never mistaken for total site size. ## Canonical access tiers ### Observer and Starter Observer provides entry-level citation readiness visibility. Starter adds more evidence detail and practical fixes for smaller teams that need the reasons behind the score. ### Alignment and Signal Alignment supports deeper diagnostics, competitor context, and recurring evidence review. Signal adds advanced validation, monitoring depth, and multi-model verification workflows. ### Agency and Score Fix Agency supports portfolio-scale workspace operations. Score Fix is a one-time remediation workflow for evidence-linked implementation changes. ## Frequently asked questions ### What is AiVIS Cite Ledger? AiVIS Cite Ledger is an AI visibility and citation readiness engine that measures whether answer engines can parse, trust, and cite a web entity with traceable evidence. ### What is BRAG? BRAG means Based-Retrieval-Auditable-Grading. It is the evidence gate that connects each finding to observable page or citation evidence and marks unsupported claims as unknown or uncited. ### What does a public report contain? A public report contains the CITE LEDGER score, platform score breakdown, key findings, prioritized recommendations, and an evidence snapshot derived from the scan record. ### How should improvements be verified? After implementing fixes, the same URL should be re-audited so new ledger rows can verify score movement, resolved blockers, and remaining citation gaps. A re-audit should confirm that the corrected page still serves the same canonical URL, includes the intended schema graph, preserves the visible answer blocks, and exposes enough policy and publisher evidence for citation engines to attribute the source without ambiguity. This closes the evidence loop from gap to fix to verified citation-readiness movement.