Website Contact & Tech Extractor — Emails, Phones, Stack

Give it a list of company domains. It visits each company's own website, follows the pages most likely to carry contact details, and returns what that company publishes about itself: email addresses, phone numbers, social profiles, VAT and SIRET numbers, and the technologies it runs.

No third-party database. No platform. No login. No terms of service bent. Just the site the domain points at — and it obeys robots.txt.

What comes back

Field	Example (apify.com)
primaryEmail, emails	support@apify.com, hello@apify.com
primaryPhone, phones	+33 4 90 82 20 22
socialProfiles, linkedin	LinkedIn, GitHub, X, YouTube, TikTok
technologies	Next.js, HubSpot, Intercom, Google Tag Manager, Cloudflare
vatNumbers, siretNumbers	CZ04788290, GB373153700
title, description	Apify: Marketplace of ready-to-run tools for AI
contactPages, pagesCrawled	/about, /contact, /contact-sales — 10 pages

primaryEmail prefers a generic address — contact@, info@, hello@, sales@ — because that is the one a company publishes to be written to.

The technology detection

Around forty signatures across CMS and e-commerce platforms (WordPress, Shopify, PrestaShop, Magento, Wix, Webflow), JavaScript frameworks (Next.js, Nuxt, React, Vue, Angular, Svelte), analytics (GA4, Matomo, Plausible, Hotjar), CRM and marketing (HubSpot, Salesforce, Klaviyo, Mailchimp), support chat (Intercom, Crisp, Zendesk, Tawk.to) and payments (Stripe, PayPal).

This is how an agency finds every prospect running an outdated stack, or how a SaaS finds everyone already using a tool it integrates with. A real run on a small French firm returned WordPress, WooCommerce, PHP/7.2 — a PHP version unsupported since 2020, which is a sales conversation on its own.

How the crawl works

Contact, about, legal-notice, impressum and privacy pages are queued ahead of everything else, in several languages, so the details are usually found within the first few pages. Static files, share links and external domains are skipped. Pages within one site are read one after another, never in parallel, while different sites run concurrently.

robots.txt is fetched first and respected. A site that disallows crawling is reported as robotsBlocked rather than silently dropped, and a site that refuses an automated request comes back with the reason — HTTP 403 — the site refused an automated request — so you can tell a blocked site from an empty one.

Where this is legal, and where the responsibility passes to you

This Actor reads pages a company published for the public to read. It touches no protected database and bypasses no access control, which is what separates it from scraping a directory or a marketplace.

What you do with the output is a different matter, and it is yours. Business contact details are still personal data under the GDPR when they identify a person — jean.dupont@firm.fr does, contact@firm.fr does not. If you prospect with them you become the data controller: you need a legitimate-interest basis, you must give the information notice at first contact, and in France you must screen against Bloctel. The Actor runs on your own Apify account and keeps nothing.

If you only need firmographic and technology data, turn extractEmails off. The technology detection, identifiers and social profiles carry no personal data at all.

Who uses this

Sales teams enriching a domain list before outreach. Agencies and web studios finding sites on an ageing stack. SaaS vendors mapping who already runs the tools they integrate with. Recruiters and researchers building a sector map. Anyone who has a list of companies and needs the way in.

Frequently asked

Why did a site return nothing? The error field says which: a refused automated request, a connection failure, a redirect to something that is not HTML, or a robots.txt that disallows crawling. An empty result is never silent.

Why so few phone numbers? Because many companies publish only a form. Numbers behind a tel: link are trusted first; loose digit runs in page text are often dates or prices and are treated with suspicion rather than reported as phones.

Can I crawl deeper? Raise maxPagesPerSite. Fifteen pages finds the contact details on almost any site; a hundred is for mapping a large site's structure.

Does it find people's names and job titles? No, and that is deliberate. This reads what a company publishes about how to reach it, not who works there.