Website Contact Extractor — emails, phone numbers and social profiles

Give it a list of websites. Get back the email addresses, phone numbers and social profiles each one publishes — as a clean table you can export to CSV, Excel, JSON, or push straight into your CRM.

It is built to run immediately after a Google Maps, directory or search scrape. Those give you a company name and a website. They do not give you a way to reach anyone. This closes that gap.

---

What it does

For every website you give it, this Actor:

1. Fetches the homepage.
2. Finds the pages most likely to carry contact details — contact, about, impressum, team, support — and fetches those too.
3. Pulls out every email address, phone number and social profile it can find.
4. Cleans the result: no image filenames mistaken for emails, no Sentry or Wix tracker addresses, no noreply@, no share-button links pretending to be the company's Facebook page.

Things most extractors miss, and this one handles

• Cloudflare-obfuscated emails (data-cfemail) — decoded, not skipped. These are addresses a human sees in the browser and a naive scraper never returns.
• Written-out addresses — sid [at] example [dot] com, sid(at)example(dot)com.
• Phone numbers in E.164 — +442079460958, not 020 7946 0958 (opt 2). Validated with Google's libphonenumber, so years, VAT numbers and prices do not come back as phones.
• Role vs personal emails — info@ and sales@ are separated from firstname.lastname@, in their own output columns, so you can route them differently.
• Expired SSL certificates — small-business sites are full of them. Handled by default instead of failing the row.

---

Input

Field	Type	Default	What it does
startUrls	array	—	The websites. Paste them, upload a CSV, or link the dataset from another Actor. Bare domains (example.com) are fine.
followContactPages	boolean	true	Also check the contact/about pages. Leaving this on roughly triples the hit rate.
maxPagesPerSite	integer	3	Homepage plus this many contact-ish pages.
defaultPhoneRegion	string	—	Two-letter code (US, GB, IN, DE) for reading local numbers written without a country code.
maxConcurrency	integer	10	Websites handled at once.
respectRobotsTxt	boolean	true	Skip pages the site asks crawlers not to fetch.
ignoreSslErrors	boolean	true	Keep going on expired or mismatched certificates.
proxyConfiguration	object	off	Optional. Most sites do not need one; switch it on if you hit blocks at volume.

Minimal input:

{
  "startUrls": [
    { "url": "https://www.apify.com" },
    { "url": "hubspot.com" }
  ]
}

---

Output

One row per website.

{
  "website": "https://decathlon.in/",
  "domain": "decathlon.in",
  "status": "ok",
  "totalContacts": 5,
  "title": "Buy Sporting Goods, Sportswear and Equipments | Download App",
  "emails": [
    "care.india@decathlon.com"
  ],
  "emailsRole": [
    "care.india@decathlon.com"
  ],
  "emailsPersonal": [],
  "phones": [],
  "socials": {
    "twitter": [
      "https://x.com/Decathlon_ind"
    ],
    "facebook": [
      "https://www.facebook.com/decathlonindia"
    ],
    "instagram": [
      "https://www.instagram.com/decathlonsportsindia"
    ],
    "youtube": [
      "https://www.youtube.com/@decathlon_india"
    ]
  },
  "pagesCrawled": [
    "https://decathlon.in/",
    "https://decathlon.in/support/hc"
  ],
  "error": null
}

status is one of:

Status	Meaning
ok	At least one contact was found.
no-contacts-found	The site was read fine and publishes nothing reachable.
failed	The site could not be read — see error (http-403, timeout, ...).
skipped-by-robots	The site's robots.txt asks crawlers not to fetch that page.

---

Pricing

Pay per website processed. One charge per website you hand in, whatever comes back — the fetching happens either way, and a site that publishes nothing costs the same work as one that publishes ten addresses. No subscription, no minimum, no charge for a run you cancel.

If a website you care about returns failed, re-run just that one with the proxy switched on.

---

Honest limits

• Only what is published. This reads the public website. It does not guess addresses, buy them from a database, or verify that an inbox accepts mail.
• No JavaScript rendering. Contact details behind a JS-only widget or a "reveal email" button will not be found. The vast majority of business sites put them in the HTML.
• Hard-blocked sites. A few sites return 403 to anything that is not a browser. Turn the proxy on for those.
• Social links are read off the site's own pages. No request is ever made to LinkedIn, Instagram or any other platform, and nothing behind a login is touched.

---

Common pairings

• Google Maps Scraper → this — turn a list of local businesses into a contactable list.
• Any directory or search scrape → this — the same last mile.
• Your own customer list → this — refresh contact details that have gone stale.

---

Compliance note

Scraping publicly published business contact details is legal in most jurisdictions, but what you do next is regulated — GDPR, CAN-SPAM, CASL and India's DPDP all govern unsolicited outreach. Personal addresses (firstname.lastname@) carry more obligations than role addresses (info@), which is exactly why this Actor separates them for you. Check your own obligations before you send.