Tech Stack Detector — website technology checker, Wappalyzer & BuiltWith alternative

A website technology checker for any list of URLs: point it at a site and get back its CMS, ecommerce platform, analytics tools, tag managers, ad pixels, CRM/marketing-automation stack, live chat widget, payment processors, CDN, hosting, JavaScript frameworks and server software — as clean, flat JSON. It reads the same kind of HTTP-response, header and HTML fingerprints as Wappalyzer and BuiltWith, using the actively-maintained, GPL-licensed webappanalyzer fingerprint database (7,628 technologies across 109 categories), and is priced at a fraction of the incumbent's rate.

• A real Wappalyzer/BuiltWith alternative, not a wrapper around one. Own detection engine (pattern matching, implies/excludes/requires chains, confidence scoring, a small DOM-subset selector matcher) reading the same fingerprint format, no scraping of either competitor.
• No browser, so it's fast and cheap. Every site is one plain HTTP fetch plus in-memory pattern matching — no headless Chrome, no per-site compute spike.
• Pay only for sites that were actually analyzed. Unreachable domains and network errors are saved as status: "error" and are never charged.
• $0.02 per site — $20 per 1,000 URLs, vs. the market-leading Wappalyzer Lookup API at $0.10/URL ($100 per 1,000).

What it detects

7,628 technologies in 109 categories. The biggest categories (technology count from the current fingerprint database):

Category	Technologies tracked
Ecommerce	803
Marketing automation	548
CMS	500
Analytics	430
Live chat	378
CRM	314
Advertising	240
Payment processors	176
Security	117
JavaScript frameworks	90
Web servers	95
CDN	69
Hosting	82
Tag managers	13

Plus WordPress/Shopify plugin & theme/app ecosystems, page builders, personalisation, appointment scheduling, reviews, and dozens of smaller categories — the full list ships in data/categories.json.

Use cases

Use case	How
B2B lead gen ("who uses tool X")	Run with no filters to get every site's full stack, then filter the dataset downstream (Sheets/Make/n8n) for rows whose technologyNames contains the tool you sell against/integrate with. Or set technologies: ["Klaviyo"] + onlyMatchingSites: true to only save (and pay for) sites that use Klaviyo at all — note a filtered run's output only lists the technologies that passed the filter, not the site's full stack, so run once unfiltered if you also need to know what CMS/ecommerce platform each lead runs on
Agency prospecting	Scan a list of local-business or industry URLs, filter categories: ["CMS", "Ecommerce"] to spot sites on outdated or low-end platforms worth pitching a rebuild to
Competitor / market research	Point it at every competitor's site, compare their categorySummary for analytics, ad pixels, CRM and marketing-automation stacks
Security / asset inventory	Scan your own domain portfolio for unmanaged CMS installs, outdated JS libraries (version field) or forgotten third-party scripts
AI agents	Flat JSON in, flat JSON out, works through the Apify MCP server — an agent can ask "what does this site run" without parsing HTML itself

Input

Field	Example	Notes
urls	["shopify.com", "wordpress.org", "stripe.com"]	One website per line/entry. A bare domain is treated as https://domain/. Duplicates (after normalization) are removed automatically.
minConfidence	50	1-100. Only keep a detected technology if its summed confidence score is at least this.
categories	["Ecommerce", "Analytics"]	Case-insensitive. Only keep technologies in one of these categories. Combines with technologies (see note above).
technologies	["Shopify", "Klaviyo"]	Case-insensitive, exact name. Only keep these specific technologies.
onlyMatchingSites	false	When true, a successfully-fetched site is only saved (and charged) if at least one technology survived the categories/technologies filters. Failed fetches are always saved as status: "error" and never charged.
includeDns	false	Also resolve TXT/MX/NS/SOA/CNAME records for the handful of fingerprints only identifiable that way. Adds a DNS lookup per site.
includeRobots	false	Also fetch and scan /robots.txt. Adds one extra request per site.
outputFormat	"site" (default) or "technology"	"site": one record per site with a nested technologies array. "technology": one dataset row per (site, technology, category) — easier to filter/pivot in Excel/Sheets. Either way you're charged once per site.
maxConcurrency	10	1-50. Sites fetched in parallel (at most one in-flight request per host regardless of this setting).
proxyConfiguration	{ "useApifyProxy": false }	Optional. Route requests through an Apify proxy if a lot of your targets block datacenter IPs. Off by default.

Example 1: basic scan

{
  "urls": ["shopify.com", "wordpress.org", "stripe.com"],
  "minConfidence": 50
}

Example 2: lead-gen list — sites using a specific marketing tool

{
  "urls": ["allbirds.com", "gymshark.com", "shopify.com"],
  "technologies": ["Klaviyo"],
  "onlyMatchingSites": true
}

Example 3: flat spreadsheet-friendly rows, one per detected technology

{
  "urls": ["allbirds.com", "gymshark.com"],
  "categories": ["Ecommerce", "Analytics", "CRM"],
  "outputFormat": "technology"
}

Output

One record per site (outputFormat: "site", default). Real output from a live run against allbirds.com:

{
  "url": "https://allbirds.com/",
  "finalUrl": "https://www.allbirds.com/",
  "domain": "www.allbirds.com",
  "status": "ok",
  "httpStatus": 200,
  "title": "Allbirds: Comfortable, Sustainable Shoes & Apparel",
  "metaDescription": "Allbirds: The world’s most comfortable shoes, flats, and clothing made with natural materials like merino wool and eucalyptus. FREE shipping & returns.",
  "language": "en-US",
  "technologies": [
    {
      "name": "Shopify",
      "slug": "shopify",
      "categories": ["Ecommerce"],
      "confidence": 100,
      "version": null,
      "website": "https://shopify.com",
      "cpe": null,
      "detectedBy": ["cookies", "meta", "scriptSrc"]
    },
    {
      "name": "Cloudflare",
      "slug": "cloudflare",
      "categories": ["CDN"],
      "confidence": 100,
      "version": null,
      "website": "https://www.cloudflare.com",
      "cpe": "cpe:2.3:a:cloudflare:cloudflare:*:*:*:*:*:*:*:*",
      "detectedBy": ["headers"]
    },
    {
      "name": "Google Tag Manager",
      "slug": "google-tag-manager",
      "categories": ["Tag managers"],
      "confidence": 100,
      "version": null,
      "website": "https://www.google.com/tagmanager",
      "cpe": null,
      "detectedBy": ["html"]
    },
    {
      "name": "Apple Pay",
      "slug": "apple-pay",
      "categories": ["Payment processors"],
      "confidence": 100,
      "version": null,
      "website": "https://www.apple.com/apple-pay",
      "cpe": null,
      "detectedBy": ["dom"]
    },
    {
      "name": "PayPal",
      "slug": "paypal",
      "categories": ["Payment processors"],
      "confidence": 100,
      "version": null,
      "website": "https://paypal.com",
      "cpe": "cpe:2.3:a:paypal:paypal:*:*:*:*:*:*:*:*",
      "detectedBy": ["dom"]
    },
    {
      "name": "Priority Hints",
      "slug": "priority-hints",
      "categories": ["Performance"],
      "confidence": 100,
      "version": null,
      "website": "https://wicg.github.io/priority-hints/",
      "cpe": null,
      "detectedBy": ["dom"]
    },
    {
      "name": "HSTS",
      "slug": "hsts",
      "categories": ["Security"],
      "confidence": 100,
      "version": null,
      "website": "https://www.rfc-editor.org/rfc/rfc6797#section-6.1",
      "cpe": null,
      "detectedBy": ["headers"]
    },
    {
      "name": "HTTP/3",
      "slug": "http-3",
      "categories": ["Miscellaneous"],
      "confidence": 100,
      "version": null,
      "website": "https://httpwg.org/",
      "cpe": null,
      "detectedBy": ["headers"]
    },
    {
      "name": "Open Graph",
      "slug": "open-graph",
      "categories": ["Miscellaneous"],
      "confidence": 100,
      "version": null,
      "website": "https://ogp.me",
      "cpe": null,
      "detectedBy": ["dom"]
    },
    {
      "name": "Swiper",
      "slug": "swiper",
      "categories": ["JavaScript libraries"],
      "confidence": 100,
      "version": null,
      "website": "https://swiperjs.com",
      "cpe": null,
      "detectedBy": ["dom"]
    }
  ],
  "technologyNames": ["Apple Pay", "Cloudflare", "Google Tag Manager", "HSTS", "HTTP/3", "Open Graph", "PayPal", "Priority Hints", "Shopify", "Swiper"],
  "categorySummary": {
    "Payment processors": ["Apple Pay", "PayPal"],
    "CDN": ["Cloudflare"],
    "Tag managers": ["Google Tag Manager"],
    "Security": ["HSTS"],
    "Miscellaneous": ["HTTP/3", "Open Graph"],
    "Performance": ["Priority Hints"],
    "Ecommerce": ["Shopify"],
    "JavaScript libraries": ["Swiper"]
  },
  "responseTimeMs": 8696,
  "error": null,
  "analyzedAt": "2026-09-18T12:40:58.254Z"
}

With outputFormat: "technology", the same detection becomes one row per (site, technology, category) — this first row is the real, unmodified output of that transform on the record above:

{
  "url": "https://allbirds.com/",
  "finalUrl": "https://www.allbirds.com/",
  "domain": "www.allbirds.com",
  "status": "ok",
  "httpStatus": 200,
  "title": "Allbirds: Comfortable, Sustainable Shoes & Apparel",
  "responseTimeMs": 8696,
  "error": null,
  "analyzedAt": "2026-09-18T12:40:58.254Z",
  "technology": "Apple Pay",
  "category": "Payment processors",
  "confidence": 100,
  "version": null,
  "detectedBy": "dom"
}

Field notes:

• detectedBy lists which signal(s) matched: headers, cookies, meta, html, scriptSrc, scripts (inline script text), url, css, dom, and (opt-in) robots/dns.
• confidence is the sum of every matching pattern's own confidence, capped at 100.
• version is null when no pattern captured a version string, not a failure.
• error and httpStatus are populated (and technologies is empty) on status: "error" rows — unreachable domains, timeouts, DNS failures, or a target the SSRF guard refused to fetch.

Pricing

Pay per event — $0.02 per site, charged once a site has been successfully fetched, analyzed, and saved:

Event	When
url-analyzed	Once per site, only when status: "ok" and the record was actually saved to the dataset

• Failed fetches (status: "error") are always saved for visibility, but never charged.
• With onlyMatchingSites: true, a successful site that has zero technologies surviving the categories/technologies filters is not saved at all — so it is not charged either.
• outputFormat: "technology" can turn one site into several dataset rows, but the charge is still once per site, not per row.
• Set Max total charge on the run to cap spend; the Actor stops cleanly (finishes in-flight fetches, no partial charges) when the cap is reached.

$20 per 1,000 sites vs. the Wappalyzer Lookup API's $100 per 1,000 (5x cheaper for the same class of HTTP-based fingerprint).

Accuracy & limits

Live-validated against 28 well-known sites with independently-known tech stacks (Shopify stores, wordpress.org, a WooCommerce store, a Magento store, HubSpot, Intercom, Cloudflare-fronted sites, vercel.com, a Nuxt site, a Drupal site, a Ghost blog, plus several JS-framework and government sites) — 16 of 20 sites with an unambiguous known platform had it correctly identified, 137 unit tests pass, and engine time is ~1.1 s/page after an indexed-DOM-matching optimization (down from 4.0 s), with peak memory 285 MB for the run.

• This is HTTP-based detection, not a browser. It reads the HTML, headers and cookies of the first response — technologies that only appear after JavaScript executes (client-side chat widgets injected by a tag manager, some JS-rendered React apps, single-page apps that hydrate content client-side) are not detected. That's the confirmed cause of every known miss in validation: Intercom's own site (its chat widget is JS-injected), react.dev (its own React usage isn't in the initial HTML), and drupal.org (only a CDN/cache header signature was visible in the raw response). ghost.org was correctly identified as not running Ghost — it actually runs Hugo on Netlify.
• One record per URL, home page only. No crawling of internal pages — pass every URL you want analyzed.
• Sites that block non-browser clients (aggressive bot detection) will return few or no signals; the Actor uses a standard desktop Chrome user agent but does not solve CAPTCHAs or render JS to get past such blocks.
• dom selector matching is a subset of Wappalyzer's dom field: tag, #id, .class, [attr], [attr="v"], [attr*="v"], [attr^="v"] and combinations of those on a single element, with exists/attributes/text checks. Multi-element combinators and JS-properties checks (which need a live DOM) are recognized and skipped, not silently guessed.
• js and xhr fingerprint fields (Wappalyzer signals that need a running browser to evaluate) are not supported — by design, since this Actor never launches a browser.

Code example

A runnable Python script (only requests needed) that calls this Actor and writes a CSV: tech-stack/find_shopify_klaviyo.py — find which sites in a list run Shopify and Klaviyo, saved to CSV.

FAQ

Is this affiliated with Wappalyzer or BuiltWith? No. This is an independent tool built on the open-source webappanalyzer fingerprint database (the maintained fork of the technologies Wappalyzer itself was originally built from), with its own detection engine.

Why didn't it detect the CMS I know a site uses? Most likely the signal only appears after JavaScript runs — see Accuracy & limits above. Try includeDns/includeRobots if the technology you're after has a DNS- or robots.txt-based fingerprint.

Why is version empty for a technology I can see the version of on the page? Only some fingerprints include a version-extracting pattern; many technologies (most CDNs, analytics tools) don't expose a version in headers/HTML at all.

Can I run this against my own internal/staging URLs? Only if they resolve to a public IP. A built-in SSRF guard refuses private, loopback, link-local, unique-local and CGNAT addresses (and re-checks on every redirect hop) so the Actor can't be pointed at internal infrastructure.

Something wrong or missing? Open an issue on the Actor's Issues tab.

License & attribution

Technology fingerprints are from enthec/webappanalyzer (GPL-3.0), the actively-maintained continuation of the original Wappalyzer fingerprint database. This Actor's fingerprint data and detection logic are distributed under GPL-3.0-only — see LICENSE. The complete corresponding source code is available to any user on request: open an issue on the Actor's Issues tab and it will be shared. Not affiliated with, endorsed by, or sponsored by Wappalyzer or BuiltWith; both names are used only to describe compatibility/comparison.