Chaos
AI Search · Google Search · Machine Intelligence

Turn search signals
into decisions that compound.

SiteNexis connects Google Search performance, AI visibility, and machine-readiness evidence in one decision layer. Observe what is changing, diagnose why, find the opportunities worth attention, and measure what improves after you act.

Start with a free site read  ·  Evidence-led, not guessed  ·  Connect Google when ready

12 intelligence scores6-stage retrieval simulationInvestigation-grade diagnostics
Scroll to discover

The Gap

Search and AI systems leave signals.
SiteNexis turns those signals into a next move.

Rankings, clicks, impressions, engagement, crawl evidence, entities, links, and machine trust all describe a different part of the same system. Looking at one score in isolation leaves the important question unanswered: what deserves attention now?

SiteNexis brings Google Search, AI retrieval, and machine-readiness evidence together. It keeps provider observations separate, shows confidence and limits, and connects every finding to an explainable action.

Screaming Frog
Ahrefs
SEMrush
PageSpeed
Schema.org
Surfer SEO

SiteNexis

A decision system for the machine-first web

Connect the evidence. Know what to do next.

SiteNexis is not another audit score. It is the operating loop between observed search performance and the conditions that make a page discoverable, understandable, trusted, and useful to machines.

01

Connect

Bind your verified site to Google Search Console and GA4.

02

Observe

Track clicks, impressions, rankings, engagement, and freshness.

03

Diagnose

Join performance to crawl, entity, truth, trust, and link evidence.

04

Find opportunities

Surface the pages and queries where verified upside exists.

05

Fix

Turn an evidence-backed opportunity into a traceable Fix Plan action.

06

Measure outcomes

Compare equivalent periods without claiming causality.

Google Intelligence

See what Google is observing about your verified property.

Connect an owned Search Console and GA4 property, then read canonical daily evidence across clicks, impressions, CTR, position, landing-page engagement, and 7/28/90-day movement.

  • Property-scoped evidence, never a guessed domain match
  • GSC query/page visibility beside legitimate GA4 page engagement
  • Freshness, completeness, provenance, and honest unavailable states
Connect Google evidence

Search Opportunity Intelligence

Find the opportunities your evidence can actually support.

Move from “your traffic changed” to a prioritized queue of striking-distance pages, high-impression/low-CTR patterns, emerging or declining queries, technical wins, internal-link opportunities, and machine-readiness gaps.

  • Deterministic detection before interpretation
  • Confidence, impact, effort, and priority kept separate
  • Every opportunity traces to Google and SiteNexis evidence
Explore search opportunities

Intelligence Modules

Machine-readiness evidence.
Connected to observed search performance.

Layer 1

SEO Intelligence

Title, meta, canonicals, robots, sitemaps, broken links, crawlability signals.

Layer 2

AI Readability

Chunk quality, entity clarity, conversational readiness, summarisability.

Layer 2

Schema Engine

Schema detection, field validation, auto-generated snippets, structured trust.

Layer 1

Link Graph

Internal link topology, PageRank distribution, orphaned pages, anchor quality.

Layer 2

Content Intelligence

Topical depth, entity coverage, citation readiness, semantic trust signals.

Layer 1

Performance

Core Web Vitals, LCP, CLS, INP, TTFB, mobile performance benchmarks.

Layer 4 — Machine TrustPro+

Retrieval Simulation · Machine Trust · Temporal Authority

Models how AI retrieval systems form, maintain, and decay trust in your content — from chunk extraction to recommendation surface presence.

Unlock Layer 4

Surface Coverage

Four AI surfaces.
One coverage map.

Retrieval and recommendation are not the same thing. A page can score well on AI Visibility and still be absent from AI Overviews, chat responses, voice answers, and agent discovery. SiteNexis maps your inclusion probability across all four surfaces — and diagnoses exactly what is blocking each one.

Google AI Overviews82
ChatGPT (browsing)67
Perplexity74
Claude (web search)91
Gemini58

All surface scores are estimated from measurable content signals.

SiteNexis OpsTelegramOperational Intelligence

Intelligence Dashboard

One scan. Twelve scores.
Complete explainability.

SiteNexis
Complete16 agents · 142 pages
74
AI Visibility+12
61
Entity Trust+8
88
Retrieval Ready+3
52
Machine Trust+21
79
Citation Score+6
Critical Issues3 critical
criticalPrimary entity definition missing across 47 pagesAI Visibility −18pts
criticalNo sameAs validation — external trust signals absentMachine Trust −15pts
warning12 pages lack structured data — retrieval riskRetrieval Score −8pts

Live AI Audit Observatory

Watch real websites
earn trust, live.

These are not mockups. Every score below comes from a real crawl, every issue from live evidence, every recommendation is deterministic. Expand a card to see the evidence, replay how it was produced, or open the full report.

Every audit above is real — run by SiteNexis against a live domain. Scores, issues, and recommendations are generated from that crawl, not a template. Run your own audit above to see the same depth of evidence.

Intelligence Scoring

Twelve scores.
Every one explainable.

Every deduction maps to a named, actionable issue. No composite score without a sub-score breakdown. No black boxes.

1
Crawl & Structure
2
Semantic Intelligence
3
AI Visibility
4
Machine Trust
View full score breakdown

example.com

Last audited 2 min ago

74
AI Visibility
+1274
Entity Trust
+861
Retrieval Ready
+388
Machine Trust
+2152
Citation Score
+679
12 issues detected3 critical

Evidence Model

A score without evidence
is just a number.

Every finding in a SiteNexis audit traces back through the same chain: evidence, an affected URL, a verification state, a confidence level, a root cause, a score impact, and a recommendation. If any link is missing, the finding does not ship.

Unavailable ≠ Zero

A category with no verified signal is excluded from scoring, never scored as a failure.

Not detected ≠ Missing

Checked-for-and-absent is a real finding, not a gap in coverage.

Crawl failure ≠ Website defect

A blocked or timed-out crawl is recorded as a measurement limit, not a site problem.

Provider failure ≠ Website failure

A third-party data source going down never becomes evidence against the site being audited.

Illustrative finding

FindingPrimary entity description conflicts between homepage and schema markup
EvidenceHomepage body text vs. Organization schema — 2 sources compared
Affected URL/about
VerificationCONFIRMED
ConfidenceHigh
Root causeEntity identity inconsistency
Score impactEntity Consistency −4 pts
RecommendationAlign schema description with homepage copy

Unavailable measurement — same finding format

FindingBacklink provider unavailable
MeasurementUnavailable
Website penaltyNone

Illustrative example — not a real audit result.

Intelligence Report v2

25 categories.
Evidence-verified, not estimated.

An additive scoring layer on top of the twelve-score report above — every one of 25 fixed categories is scored only from evidence that was actually verified, with coverage and confidence reported as separate, honest fields.

Why this score?

Evidence
Findings
Contradictions
Root Causes
Scoring
Recommendations
Opportunities

Categories with no verifiable evidence are marked UNAVAILABLE and excluded from the weighted total — the overall score is normalized only across what was actually measured.

Read the full category breakdown

Sample audit

Illustrative — not a real customer result

64/100Grade: Fair

Coverage

82%

Confidence

Moderate

Discoverability & Technical
71
AI Visibility
58
Identity
64
Authority & Citation
49
Trust & Governance
77

Methodology

How SiteNexis works.
Every step, every agent, every score.

01

Full-Site Crawl

The Crawl Agent renders every page with Puppeteer, strips navigation and boilerplate, extracts semantic chunks of 300–600 tokens, and pre-identifies entities. Up to 500 pages per audit. Respects robots.txt.

02

Parallel Layer 1–2 Analysis

Six agents run simultaneously: SEO (title, meta, canonicals, sitemaps), Schema (structured data validation), AI Readability (chunk quality, extractability), Entity Intelligence (consistency, disambiguation), Citation Probability (7-factor weighted score), and Semantic Trust (contradiction detection via Claude API on top 20 pages).

03

Layer 3 AI Visibility Scoring

AI Visibility Score = Machine Readability × 15% + Entity Confidence × 20% + Retrieval Readiness × 20% + Citation Probability × 20% + Semantic Trust × 15% + Schema Completeness × 10%. The AI Perception Graph is built from entity relationships and typed connections. Every deduction maps to a named, actionable Issue.

04

Layer 4 Machine Trust (Pro+)

Five Layer 4 agents run in parallel after Layers 1–3 complete. Retrieval Simulation models 6 stages on top 30 pages by PageRank. Machine Trust scores entity credibility, schema alignment, external validation, contradiction absence, and decay signals. Temporal Authority tracks velocity and drift across audits. Recommendation Mapping scores inclusion probability across four AI surfaces. Synthetic Entity Detection identifies manufactured authority patterns.

05

Information Gain + Scout

The Information Gain Engine fetches the SERP cohort via Serper API and measures what your content adds that the top results do not already cover — entity gaps, question gaps, evidence uniqueness. Scout classifies every page by query intent (informational, commercial, research, voice, agent) and scores structural alignment between intent and page signals.

06

Global Fix Plan + Reports

The Fix Plan aggregates every issue from all 12 modules into a single P0/P1/P2 queue with cross-module dependency chains, effort estimation, and three-dimension impact scoring (SEO / AI Visibility / Trust). The Reporting Agent generates a full PDF with all scores, issues, and recommendations. Every API route returns a GTL state envelope: complete, partial, or empty.

Pricing

Start free.
Scale with intelligence.

Plans unlock deeper evidence and more properties. They do not promise rankings, citations, or outcomes; they give your team a clearer measurement loop and more room to act.

Free

$0

A clear starting point for your first site.

  • 1 site intelligence read / month
  • Core crawl + machine-readiness evidence
  • Evidence-led report
  • No credit card
Start Free
Starter

$29/mo

For teams building a repeatable search loop.

  • 50 site intelligence reads / month
  • Google Intelligence foundations
  • Search and AI evidence reports
  • Email support
Get Started
ProMost Popular

$79/mo

For teams acting on verified opportunities.

  • Unlimited site intelligence reads
  • Google Intelligence + historical evidence
  • Search Opportunity Intelligence
  • Machine Trust and Fix Plan context
  • Priority support
Start Trial
Agency

$249/mo

For agencies managing multiple properties.

  • Unlimited properties and intelligence reads
  • Bulk domain scanning
  • White-label reports
  • API access
  • Dedicated manager
Contact Sales

Product Suite

Two platforms.
One intelligence suite.

SiteNexis models how AI systems retrieve and trust your content. AdNexis models what creative drives conversion when they do. Together, they cover the full signal-to-conversion pipeline.

SiteNexis

AI Retrieval & Machine Trust Intelligence

You are here

Models how AI systems retrieve, interpret, trust, and recommend your website — across all four intelligence layers from crawl structure to machine trust formation.

  • Machine Trust Score — 5-dimension trust model
  • Entity Intelligence — consistency + disambiguation
  • Retrieval Simulation — 6-stage pipeline modeling
  • Recommendation Surfaces — AI Overviews, chat, voice, agents

AdNexis

AI Creative Intelligence

New

Deconstructs top-performing ads with AI to extract what makes them convert — then generates platform-specific creative variations in seconds.

  • Hook Intelligence — classify, score, and rank ad hooks
  • Full Ad Analysis — funnel stage, CTA, emotion, audience
  • Swipe Vault — your AI-analysed ad library
  • Creative Generation — variations for Meta, TikTok, YouTube

Machine Resource Studio

Turn crawl data into clear decisions.

Machine Resource Studio explains what the crawl found, why the finding matters to Google and AI systems, and which fix should come first. Each score shows its formula, evidence count, confidence, and limits.

Explore the studio

Evidence first

Recommendations link to crawl pages or issue records.

Explainable scores

Every score card shows the formula and confidence.

Google and AI views

Read technical impact and machine interpretation impact separately.

Honest limits

Missing data stays unavailable. It does not become zero.

Operational Intelligence

Built to watch the web.
Built to watch itself.

SiteNexis Ops is internal operational infrastructure — it watches SiteNexis's own audit pipeline, provider availability, and platform health, and raises alerts to a private Telegram channel when something needs attention. It is not a customer-facing chatbot.

AI / Search Ecosystem
SiteNexis
SiteNexis Ops
Telegram
Operational Intelligence

Monitors: audit failures · partial/stalled audits · abnormal audit performance · provider degradation · integration health · platform health · operational incidents.

SiteNexis Ops

Provider degradation detected

Scout enrichmentRate limited
FallbackDeterministic classifier active
Core auditContinuing normally
ImpactLimited

Illustrative example — not a live incident. Provider unavailable ≠ website failed. Fallback active ≠ audit failed.

Get Started Free

See what search and AI systems
actually extract from your site.

Connect one site, observe the evidence, and find the next supported action. Start with a free site read; connect Google when you want property-scoped performance history.

Google Intelligence and Search Opportunity Intelligence require an authorized, verified property binding.