How AGNURA Detects Website Platforms and Themes

AGNURA’s platform and theme labels are best-guess heuristics from the page you have open. They are useful for outreach research. They are not certificates that prove how a site was built.

Higher confidence means more overlapping signals matched. It does not mean the guess is always right, especially on hybrid or heavily customized sites.

Hub context: What Is Website Intelligence for Sales and Outreach?. How to run analysis: Run Your First Website Analysis.

Direct answer

Output	What it is	What it is not
Platform name	Best-scoring stack guess (or Unknown)	A vendor audit or BuiltWith replacement
Confidence 0–100	Weighted signal score, capped at 100	Proof or probability from a lab test
Signals list	Short reasons that matched (truncated in UI)	Full forensic evidence
Theme	Best-effort Shopify or WordPress theme name/slug when detectable	Guaranteed theme identity on every stack

Always open the live site (and View Source when unsure) before you treat the label as a talking point in outreach.

Why heuristics are used

Sales research needs a fast answer: “Is this roughly Shopify or WordPress?” Dedicated tech databases help at scale. AGNURA’s job here is active-tab speed: read the DOM and assets you already loaded, score common fingerprints, show the guess with reasons.

That design stays local to the page analysis path (no managed AI for platform detection). Network probes like /sitemap.xml are separate and can fail while platform/theme still look complete offline for the already-loaded HTML.

How platform detection works

AGNURA’s content script scores candidate names and picks the highest total.

Signal families

Family	Examples
Meta generator	meta[name="generator"] mentioning WordPress, Shopify, Wix, Squarespace, Webflow, Next.js
Asset URLs	Paths/hosts such as wp-content, cdn.shopify.com, Wix/Squarespace/Webflow CDNs, BigCommerce, WooCommerce, /_next/, Demandware/SFCC patterns
Window globals	e.g. Shopify, __NEXT_DATA__, some Wix runtime globals (when accessible)
Body / HTML hints	e.g. WooCommerce body class; myshopify.com in HTML; WooCommerce plugin asset paths

Weights differ by signal strength (generator and strong globals count more than a weak HTML mention). Confidence is roughly the winning score, capped at 100. The UI may show a truncated signals list (implementation keeps a limited set).

Example labels the scorer knows

Depending on what matches, the best-guess name can be things like WordPress, Shopify, Wix, Squarespace, Webflow, BigCommerce, WooCommerce, Next.js, Salesforce Commerce Cloud, or Unknown.

WooCommerce can appear alongside WordPress signals on the same site. The winner is whoever scores highest on that page load, not a full dependency graph.

How theme detection works

Theme detection is stack-aware and mainly useful for Shopify and WordPress.

Shopify

Theme name may come from:

• window.Shopify.theme when the content script can see it
• Inline script text that contains Shopify.theme (common when the isolated world cannot see main-world window details)

If Shopify assets are present but no theme name is found, you may get a “theme not found” style result with explanatory signals.

WordPress

Theme slug/name may come from:

• /wp-content/themes/{slug}/ on stylesheet/script URLs
• body classes like wp-theme-{slug} (helpful for some block themes)

Slugs are humanized for display when possible. If WordPress-style assets exist but no theme path matches, the theme field can stay empty or note that no slug matched.

Other platforms

Many stacks will show platform without a meaningful theme. That is expected.

Options and trade-offs

Approach	Upside	Trade-off
AGNURA active-tab heuristics	Fast; explainable signals on the open page	Can mislabel hybrids; not a full internet fingerprint DB
Dedicated tech lookup tools	Broader historical coverage	Different product job; may lag live DOM
Manual View Source	Highest control	Slower
Trusting the label blindly in email	Fastest	Highest embarrassment risk

Decision framework

If…	Then…	Because…
Confidence is high and signals look coherent	Use as a research hint	Multiple fingerprints aligned
Confidence is middling or signals conflict	Verify manually	Hybrids and CDNs confuse scorers
Platform is Unknown	Do not invent a stack in outreach	Unknown means no winning pattern
Theme is missing on Shopify/WordPress	Check theme assets yourself or skip the claim	Detection is best-effort
You will mention the stack in a cold email	Confirm on the live storefront/admin clues you can see	Heuristic ≠ proof

Checklist: read platform/theme safely

1. Run analysis on the real public page that matters (not a login wall).
2. Note platform name + confidence.
3. Expand or read the signals when shown.
4. Note theme only if present and you can spot matching assets.
5. For outreach, convert “Shopify (heuristic)” into language you can defend, or omit it.
6. Reload analysis if the page was still hydrating when you first opened the popup.

Risks and limits

• Custom fronts on Shopify/WordPress can hide or fake fingerprints.
• Next.js or other frameworks can sit under a CMS; labels reflect what scored highest on this page.
• CDN and tag-manager URLs can add noise.
• This is not legal proof of vendor relationship or license.
• Platform labels can change product-count merge behavior (Shopify product sitemaps vs WordPress post sitemaps). That is a separate metric with its own limits: What a Sitemap Page Count Can and Cannot Tell You.

How AGNURA uses these guesses

Downstream use	Role of platform/theme
Popup research UI	Hero technology icon and labels
Email templates	Optional placeholders like {{platform}} / {{platformConfidence}} (still verify)
Generate AI Email context	May include platform in the analysis bundle
Product count merge	Prefer Shopify vs WordPress sitemap rollups when labels match
SEO Analysis entry	Opens under the platform/tech icon (audit itself is rules-based, not AI)

Compared with common alternatives

Approach	Fit
BuiltWith-style databases	Scale and history
Wappalyzer-style extensions	Similar heuristic family
AGNURA platform/theme	Same-tab research tied to outreach workflow

Related guides

• What Is Website Intelligence for Sales and Outreach?
• How to Research a Prospect’s Website Before Writing an Email
• What a Sitemap Page Count Can and Cannot Tell You
• Run Your First Website Analysis
• How to Personalize an Outreach Email From Website Research

Soft next step

Analyze one known Shopify store and one known WordPress site, compare confidence and signals, and notice where theme appears or does not. Install AGNURA if you want that in the popup:

Add to Chrome