NovaX is a site-wide AI visibility intelligence platform. While other tools track how many times your brand appears in AI responses (outcome monitoring), NovaX analyses the actual signals that make AI systems cite a page — structural integrity, entity clarity, schema markup, semantic richness, and AI crawler access.
NovaX works in two layers:
Server crawl layer
NovaX crawls your entire site automatically, discovering every page, detecting schema markup, checking AI bot access via robots.txt, and mapping internal links. This gives you a baseline picture of every URL on your domain.
Inspector deep layer
When AI Visibility Inspector runs on a page, it performs deep DOM analysis — detecting entities, scoring semantic richness, calculating per-engine compatibility, and identifying query eligibility. This data flows silently into NovaX, upgrading the page from crawl-only to full signal depth.
The key insight: Every competitor tracks what AI says about you. NovaX is the only tool that tells you why AI reads your pages the way it does — and exactly what to fix.
Metric cards (top row)
Four key numbers: total pages crawled, Inspector-fresh pages (deep data), crawled-only pages (need Inspector), and AI bots blocked. The color-coded top border indicates health — blue/teal is good, amber is attention needed, red is urgent.
AI Visibility Score ring
The site-wide average AI visibility score (0–100), with a breakdown of the five signal dimensions below it. The ring fills from blue to teal as your score improves. Only populates when Inspector has run on pages.
Page status bars
Three bars showing the split between Crawled only (red), Inspector stale (amber), and Inspector fresh (teal). The goal is to move all bars to teal by running Inspector regularly.
Work queue
The ten most urgent pages needing attention — sorted by tier and score. Click any row to open the full page detail. This is your Monday morning action list.
AI crawler access
Shows whether GPTBot, ClaudeBot, PerplexityBot and Google-Extended can access your pages. Green tick = allowed. Red X = blocked. Dash = no explicit rule in robots.txt.
Engine compatibility
Average scores for each AI engine — Perplexity, ChatGPT, Claude, Gemini. Shows average citation probability per engine. Populated by Inspector data.
The page inventory shows every URL discovered during the last crawl. Use the stat cards at the top to filter by status, or use the search box to find any specific page or title.
Status filters
Filter by Crawled only, Inspector stale, Inspector fresh, Tier 1 or Tier 2. The stat cards are also clickable filters. Red badge on the nav shows how many pages still need Inspector.
Page speed column
Load time detected during crawl. Green = under 1s, amber = 1–3s, red = over 3s. Slow pages reduce AI crawlability scores.
Word count column
Total words detected in the page body. Thin content (under 300 words) typically scores poorly on semantic richness.
Click any row
Clicking any page row opens the full page detail view with all available signal data, recommendations, entities, schema status and link counts.
The page detail view opens when you click any page in the inventory or work queue. It shows all available data for that specific URL.
Page hero
URL, page title, status pill, tier badge, HTTP status code, crawl time, Inspector time, page speed and word count. The teal banner appears when Inspector data is missing, inviting you to run it.
Signal breakdown
Five horizontal bars showing individual signal scores — Structural Integrity, Data Extractability, Entity Clarity, Schema & Metadata, Freshness & Decay. Only available after Inspector has run on this page.
Detected entities
Named entities Inspector found on this page — products, technologies, concepts, standards. Primary entities (strong signal) shown in blue, secondary in grey, weak in dark grey.
Recommendations
Prioritised list of fixes — high (red border), medium (amber), low (blue). Each recommendation has a title and description. Resolved automatically when Inspector re-runs and detects improvement.
| Signal | What it measures | How to improve it |
| Structural integrity | Heading hierarchy (H1–H4), logical content flow, HTML cleanliness, mobile responsiveness signals | Ensure one H1 per page, logical heading nesting, clean semantic HTML structure |
| Data extractability | How easily AI can parse and extract information from your page — table structure, list formatting, clear paragraph breaks | Use tables for comparative data, bullet lists for features, short focused paragraphs |
| Entity clarity | Named entities present and their Knowledge Graph connection — products, technologies, concepts, standards, people, organisations | Name things explicitly, use industry standard terminology, reference known entities by their proper names |
| Schema & metadata | JSON-LD structured data presence and completeness — Article, Product, FAQ, HowTo, BreadcrumbList and other schema types | Add appropriate schema markup for your page type, complete all recommended fields |
| Freshness & decay | Content recency signals — publication date, last modified date, date references in content, update frequency | Add explicit publication and update dates, refresh older content, reference current year where relevant |
85
80–100 — Excellent. This signal is a genuine strength. AI engines can read, understand and cite this content confidently.
68
60–79 — Good. Solid signal. Minor improvements would push this into excellent territory.
48
40–59 — Moderate. Detectable weakness. AI engines can access and read this content but miss important signals.
28
0–39 — Critical. Significant gap. This signal is actively reducing your AI citation likelihood.
Crawled only
NovaX has crawled this page and detected basic data (title, speed, word count, schema, bot access) but Inspector has never run on it. Signal scores are not available. Action: open this page in your browser and run Inspector.
Inspector stale
Inspector has run on this page but the data is older than 30 days. Scores may not reflect current page content. Action: re-run Inspector on this page to refresh the data.
Inspector fresh
Inspector has run on this page within the last 30 days. All signal scores, entities, recommendations and query eligibility data are current and reliable.
The signal heatmap shows every page in your site as a row, with each signal dimension as a colour-coded cell. Teal = strong (80+), blue = good (60–79), amber = moderate (40–59), red = critical (below 40), dash = no data yet.
How to use it: Look for entire columns that are red or amber — that's a systemic gap. If the Schema column is red across all your pages, one structured data sprint fixes your biggest weakness site-wide. If only specific rows are red, those are individual page problems.
The ● / ○ column shows Inspector verified vs crawl-only pages. Verified scores are exact; crawl-only scores are estimates. Use the filter buttons to focus on Tier 1 pages or Inspector-verified pages only.
Inspector and NovaX communicate automatically once connected. When Inspector runs a full analysis on any page, it silently POSTs the complete signal data to NovaX — no extra steps needed.
API endpoint
The NovaX API endpoint is shown in Workspace Settings. Inspector sends a POST request to this URL with the full analysis payload whenever it completes a page scan.
What gets stored
Overall score, all five signal dimensions, per-engine scores, citation probabilities, detected entities, schema markup validation, query eligibility data, and prioritised recommendations.
Status upgrade
The moment Inspector data arrives for a page, its status automatically changes from "Crawled only" (red) to "Inspector fresh" (teal). The dashboard score ring and signal heatmap update immediately.
Historical tracking
Every Inspector run is stored separately with a timestamp. This builds the score history that powers the Score history screen — showing month-by-month progress as you improve your pages.
Pro tip: Run Inspector on your highest-traffic Tier 1 pages first. Those scores appear immediately in the dashboard and give you the most actionable data fastest.
1
Run the crawler monthly
Start each month with a fresh crawl to discover new pages, detect any robots.txt changes, and update page metadata. The crawler runs in 2–5 minutes depending on site size.
→ Run crawler
2
Check the signal heatmap
Open the heatmap and look for red columns — systematic weaknesses. A red Schema column across all pages means one JSON-LD sprint fixes your biggest site-wide gap.
→ Signal heatmap
3
Work through the queue
Open the dashboard work queue and start with the lowest-scoring Tier 1 pages. Open each page in your browser, run Inspector, fix the top recommendations, run Inspector again to confirm improvement.
→ Dashboard work queue
4
Fix AI bot access issues
Check the AI Bot Access screen for any blocked or no-rule pages. Update robots.txt to explicitly allow GPTBot, ClaudeBot, PerplexityBot and Google-Extended.
→ AI bot access
5
Address schema gaps
Check the Structured Data screen for pages missing schema. Add appropriate JSON-LD markup — Article for blog posts, Product for product pages, FAQ for FAQ sections, HowTo for guides.
→ Structured data
6
Track progress in Score history
After a month of improvements, the Score history screen shows your trend. Share this with clients or management as proof of AI visibility improvement.
→ Score history
Each workspace is a fully isolated environment for one website — its own pages, scores, crawl history, alerts, API keys, and client logins. One NovaX installation can manage multiple workspaces up to your licence plan limit.
Workspace selector — the landing page at /workspaces.html shows all your sites as cards. Click any card to enter that workspace. The licence bar shows your plan, expiry, and how many sites you have used vs your limit.
Workspace switcher — click your domain name at the top of the sidebar on any screen to open a dropdown. Switch between workspaces instantly without going back to the selector.
Add New Site — click the dashed card on the workspace selector. Enter name and domain. NovaX creates the workspace and generates an API key automatically. The button is disabled when you reach your plan's site limit.
Licence — one key covers your entire installation. Validate or update it in Settings → Licence. Your plan (Solo / Professional / Agency / Enterprise) determines how many sites you can add.
The Pre-Publication Optimizer analyses a content draft against all five AI visibility signal dimensions and tells you exactly what to fix before publishing. No URL needed — just paste the text.
How to use — paste your draft from Word, Google Docs, WordPress, or any text editor into the text area. Optionally enter a target topic and audience. Click Analyze draft for instant results.
What it detects — heading structure, word count, bullet lists, tables, named entities, author/date signals, FAQ sections, freshness language, citations, and schema opportunities.
Benchmark — if you are in a workspace with existing Inspector data, the optimizer compares your draft scores against your site average so you know whether the draft is above or below your existing content.
Fix before publishing — each issue is listed with a priority (high / medium / low), the signal it affects, and a specific action step. Fix the high-priority items before publishing for maximum AI visibility impact.
→ Pre-pub optimizer
When Inspector syncs a page and detects a score drop of 5 or more points vs the previous run, NovaX automatically creates an alert and sends an email to the workspace alert address.
Setting your alert email — go to Settings → Workspace Details → Alert email. Enter your email address (or your client's) and click Save changes. Each workspace has its own alert email.
Email content — the alert email shows the page URL, before/after score, the drop amount, and a direct link to the page detail in NovaX.
Dashboard alerts — unread alerts also appear as a red panel at the top of the dashboard. Click any row to open the page detail. Click "Dismiss all" to clear them.
What to do — open the page detail, check which signal dimension dropped, read the Recommendations section, fix the issue on the live page, re-run Inspector to confirm recovery.
NovaX checks every hour whether any workspace is due for a crawl based on its crawl frequency setting. When due, it fires the crawl automatically in the background.
Setting frequency — go to Settings → Workspace Details → Crawl frequency. Options: Daily (every 24h), Weekly (every 7 days), Monthly (every 30 days). Set to Manual to disable auto-crawl for that workspace.
How it works — a cPanel cron job runs every hour and calls the NovaX scheduler. The scheduler checks each workspace, compares the last crawl date against the frequency, and fires if due.
Manual crawl — you can always run a manual crawl from Run crawler in the sidebar, regardless of the scheduled frequency. Choose full or incremental mode.
Deleting old jobs — crawl history in Settings can grow long. Click the red trash icon next to any completed or failed job to delete it. Running jobs cannot be deleted.
→ Run crawler
The client portal gives your clients a clean, read-only view of their workspace data — scores, signals, page inventory, and alerts — without access to any settings, crawler controls, or API keys.
Adding a client login — go to Settings → Client Access. Enter the client's name, email, and a password. Click Add client. Their login is tied to the current workspace only.
Portal URL — send clients to /portal-login.html. The URL is shown in Settings → Client Access with a Copy button. Clients log in with their email and password.
What clients see — AI Visibility Score ring, five signal bars, page count stats, score drop alerts, and the full page inventory with individual scores per signal.
Managing clients — disable a client login temporarily using the Disable button, or remove it permanently with Remove. Clients are workspace-specific — a client on one workspace cannot see another.
White-label — if white-label is enabled, clients see your agency branding instead of NovaX. See the White-label section below.
Agencies can replace all NovaX branding with their own name, logo, and colours. Clients see the agency's brand — "NovaX" is never visible to them.
What can be changed — company name, tagline, sidebar logo (image or letter mark), primary accent colour, background colour, browser tab title, and support email shown in Help.
How to apply — edit white-label.js in the server root. Set enabled: true and fill in your branding details. Restart the Node.js app. The entire platform reflects your branding instantly.
Logo — set logoUrl to an image URL to replace the letter mark with your logo. Set logoHeight to control the size (e.g. '28px').
Colour — set colorPrimary to your brand hex colour (e.g. '#e63946'). This changes all teal accents, active nav items, score rings, and buttons throughout the platform.
The Citation Tracker runs live queries against AI engines (Perplexity, ChatGPT, Claude, Gemini) and checks whether your pages appear in the answers. It shows citation probability, which queries trigger citations, and how your visibility changes over time.
How it works — enter a query or pick from your tracked queries. NovaX submits it to each AI engine and analyses the response for mentions of your domain, page titles, or entity names.
Citation probability — shown as a percentage per engine. A score of 70% means 7 out of 10 similar queries returned your content in the answer.
Query limits — select 10 to 1,000 queries per run. Higher counts give more reliable data but take longer and use more API credits.
History — every run is stored. Track citation trends over time to see whether content improvements translate to higher AI visibility.
The Classic SEO module surfaces traditional search signals — meta data, heading structure, internal linking, word count, and on-page optimisation — alongside your AI visibility scores so you can manage both in one place.
On-page signals — title tag, meta description, H1, canonical URL, word count, and internal link count per page.
Coverage — all crawled pages appear here. Filter by missing titles, missing descriptions, thin content, or duplicate signals.
Relationship to AI visibility — strong classic SEO signals (clear titles, structured headings, schema) directly improve AI retrievability. Fixing classic SEO issues improves both rankings and citations.
Batch scan queues all crawled pages and drives the Inspector extension through them automatically — one by one — syncing results back to NovaX as it goes. No manual clicking required.
How to start — go to the Dashboard, click "Batch scan", choose a mode, and confirm. The Inspector extension picks up the job within about 30–60 seconds and begins.
Modes — Crawled only scans pages with no Inspector data. Stale rescans pages not updated in 7+ days. All rescans every page regardless of freshness.
Requirements — the Inspector extension must be installed and connected to NovaX (URL + workspace ID set in its settings). It runs entirely in the background — you do not need to keep the popup, a tab, or the browser window open or active. It will keep working even if the browser is minimized.
Progress — a live counter on the dashboard shows pages completed vs total, and the extension's toolbar icon shows a running count too. You can cancel at any time.
Resilience — if Chrome restarts or closes the extension mid-scan, it automatically resumes from the exact page it left off on within about a minute. No progress is lost.
The Users module lets administrators invite team members, assign roles, and control who can access which workspaces. Each user has their own login — no shared passwords.
Roles — Admin has full access including user management and settings. Editor can run scans and view all data but cannot change settings or manage users. Viewer has read-only access.
Inviting users — go to Users, click "Invite user", enter their email and role. They receive a login link by email.
Workspace access — users can be scoped to specific workspaces. An editor for Client A cannot see Client B's data.