Tech Stack Detector — website technology lookup, with the evidence attached

Find out what any website is built on. Frameworks, CDNs, ecommerce platforms, CMSs, analytics, chat widgets, payment providers — from one polite request per site.

Quick start: press Start with no configuration at all. It ships with twelve well-known sites so a single run shows you the whole output shape. Then paste in your own list.

Sample output

url	title	technologyCount	categories	server
https://shopify.com/	Shopify	2	CDN, Ecommerce	cloudflare
https://vercel.com/	Vercel	2	Framework, Hosting	Vercel
https://wordpress.org/	WordPress.org	3	Analytics, CMS, Web server	nginx
https://gitlab.com/	GitLab	3	CDN, Framework, Marketing	cloudflare

Every row also carries target, finalUrl, httpStatus, reachable, robotsAllowed, robotsReason, description, poweredBy, checkedAt, error, and a technologies array where each entry has a name, a category and an evidence string.

Why this one

Every detection carries its evidence. A bare list of technology names is unfalsifiable — you either trust it or you do not, and you have no way to tell which. Here each entry says exactly what matched:

{ "name": "Shopify", "category": "Ecommerce",
  "evidence": "page references cdn.shopify.com/shopifycloud/..." }

You can check it. You can argue with it. And when a signature is wrong, you can see why it was wrong instead of quietly building a campaign on it.

A site's own claim is labelled as its own claim. The <meta name="generator"> tag is the one signal a site can set to anything at all, so anything found only there is filed under the category Self-declared rather than mixed in with things we actually detected.

Nothing matched means nothing matched. A plain HTML page returns an empty list, not a hopeful guess. The signatures match real markers — /_next/static/, cdn.shopify.com, a cf-ray header — never the vendor's name in prose, so a careers page that says "we use React" does not become a React detection.

It reads the site's robots.txt first. You named the target, so this is not crawling — but the Actor still checks and skips any path the site asks automated clients to leave alone, and says so in robotsAllowed and robotsReason rather than doing it silently. A skipped fetch and a site with no technology are very different answers.

What you can do with it

• Qualify leads before you write to them. Knowing a prospect runs Shopify, HubSpot and Klaviyo tells you what to sell and what to say.
• Size a market. Run a list of a thousand competitors' customers and count how many are on the platform you integrate with.
• Watch a competitor's stack change. Schedule it and diff technologies over time; a new analytics vendor or a CMS migration is a visible event.
• Audit your own estate. Point it at every domain you own and find the forgotten WordPress install.
• Feed an AI agent. Rows are around 1 KB, identically shaped, and every claim is accompanied by its justification, which is exactly what a model needs to reason rather than parrot.

Input

Field	Default	Notes
websites	12 sample sites	A bare domain or a full URL. A URL with a path analyses that page, not the home page.
respectRobots	true	Skips paths the site's robots.txt disallows, and reports it. Turn off only for sites you own.
onlyWithTechnologies	false	Drops rows where nothing was detected. You are charged per row returned, so this is cheaper on long lists.
concurrency	5	Parallel fetches. One request per site either way.

Limits worth knowing before you buy

This reads the HTML the server returns; it does not run JavaScript. A technology loaded only after the page boots in a browser will not be seen. That keeps the Actor fast and cheap — around a second per site — and it means the answer is "what the page declares", not "everything running in a session".

Detection is signature-based, so it finds what it has a signature for. Around sixty technologies across hosting, CDN, frameworks, CMS, ecommerce, analytics, marketing, support, payments and monitoring — the ones that matter for qualifying a company, not an exhaustive registry.

Sites behind an aggressive bot wall will answer with a challenge page. That is reported as the HTTP status it was, not silently turned into an empty result.

Pricing

Pay per site returned. Sites removed by your filters are not charged, and duplicates are dropped before anything is counted.

Disclaimer

This Actor fetches pages you name, one request each, identifying itself honestly in its user agent. It scrapes no website catalogue and calls no third-party API — there is nothing between you and the site you asked about.

It reads each target's robots.txt and honours it by default. Technology names are trademarks of their respective owners and are used here only to describe what a page appears to be built with; this Actor is not affiliated with, endorsed by or connected to any of them.

Detection is inference from public evidence, not certainty. The evidence field exists so you can judge each conclusion yourself rather than take it on trust.