📧 Email Scraper & Contact Finder - Bulk Emails, No API Key avatar

📧 Email Scraper & Contact Finder - Bulk Emails, No API Key

Pricing

from $6.00 / 1,000 results

Go to Apify Store
📧 Email Scraper & Contact Finder - Bulk Emails, No API Key

📧 Email Scraper & Contact Finder - Bulk Emails, No API Key

Email scraper and contact finder: pull bulk emails, phones and social links from any website with no API key. Smart crawler hits contact, about and team pages first for fast, high-hit-rate results, and detects the tech stack too. No login. Ideal for B2B lead gen. Pay per website crawled.

Pricing

from $6.00 / 1,000 results

Rating

0.0

(0)

Developer

Renzo Madueno

Renzo Madueno

Maintained by Community

Actor stats

1

Bookmarked

247

Total users

3

Monthly active users

3 days ago

Last modified

Features

Limited permissions

Share

📧 Email Scraper & Contact Finder for Any Website

Give it a list of websites and it reads each one (home page first, then the contact, about, team and imprint pages) and returns one row per website: emails, phone numbers, postal addresses, social profiles, contact page URL and tech stack. Built for lead generation, outbound sales and CRM enrichment.

Pricing: pay per result row plus a small run-start fee. The current rates are on the Pricing tab. No subscription, no API key.

What you get per website

FieldDescription
statusfound (contact details were found), no-contacts-found (the website was read and publishes none), and, only if you ask for them, blocked or unreachable
emailsThe addresses the website publishes for reaching it, on the pages read (up to 50): addresses on its own domain, addresses behind a mail link or on its contact, about, team and legal pages, and free-mail addresses such as Gmail. An address of another company that only appears in running text is not listed here; it stays in emailDetails
emailDetailsThe same addresses with context: type (role for shared mailboxes such as info@ or sales@, personal for addresses shaped like first.last@, other when the address alone does not tell), onCompanyDomain, foundIn (mailto, protected, structured-data, page, spelled-out) and the sourceUrl where each one was seen. Addresses on the company's own domain come first
phonesPhone numbers in international format (+1 888 926 2289). Numbers the website publishes as call links or in its structured data are preferred; numbers read from page text are used only when there are none, and phonesVerifiedByLink tells you which case applies. A number found only in page text with neither a + prefix nor a label such as "Phone" or "Tel" next to it is listed, but does not by itself make the row found
addressesPostal addresses the website declares in its structured data
socialLinksLinkedIn, Twitter/X, Facebook, Instagram, YouTube, TikTok, GitHub, Pinterest, Telegram, WhatsApp and Threads profiles
contactPageUrlThe website's contact page, if one was found among the pages read
hasContactFormtrue when one of the pages read has a form with a message box, for companies that publish no address. Newsletter sign-up boxes do not count
companyNameTaken from the website's own metadata, its title or its domain
techStackWhich of 24 platforms and tools the website runs on
methodhttp or browser: how the website had to be read
pagesCrawled / scrapedAtPages read and timestamp
detailFor blocked and unreachable rows, the reason

Example

Input

{
"domains": ["stripe.com", "https://basecamp.com"],
"maxPagesPerDomain": 4
}

Output (a real row from a run on 10 November 2026)

{
"domain": "stripe.com",
"companyName": "Stripe",
"emails": ["sales@stripe.com"],
"emailDetails": [
{
"email": "sales@stripe.com",
"type": "role",
"onCompanyDomain": true,
"foundIn": "page",
"sourceUrl": "https://stripe.com/contact/sales"
}
],
"phones": ["+1 888 926 2289"],
"phonesVerifiedByLink": false,
"addresses": [
"60 Anson Road #05-01, Singapore, 079914, SG",
"920 5th Avenue, Suite 1900, Seattle, WA, 98104, US",
"201 Bishopsgate, 5th Floor, London, EC2M 3AB, GB",
"350 N. Orleans, Suite 5000S, Chicago, IL, 60654, US",
"28 Liberty Street, 48th floor, New York, NY, 10005, US"
],
"socialLinks": {
"linkedin": "https://www.linkedin.com/company/stripe",
"twitter": "https://twitter.com/stripe",
"facebook": "https://www.facebook.com/StripeHQ",
"instagram": "https://www.instagram.com/stripehq",
"youtube": "https://youtube.com/@stripe",
"github": "https://github.com/stripe"
},
"contactPageUrl": "https://stripe.com/contact/sales",
"hasContactForm": false,
"techStack": ["Next.js", "React", "Stripe"],
"pagesCrawled": 4,
"method": "http",
"scrapedAt": "2026-11-10T22:57:36.858Z",
"status": "found"
}

Which websites get a row, and what you pay for

You pay for rows, so it is worth knowing exactly which websites produce one.

  • One row per website, however many pages were read to build it. The same website given twice, or once as a URL and once as a domain, is read once. A URL is reduced to its host: https://example.com/about and example.com are the same website, and the domain column holds the host in lower case.
  • A website that was read gets a row, with status found or no-contacts-found. Turn on onlyWithContacts if you want rows only for websites with contact details; the others are then listed in the SUMMARY record and not charged.
  • A website that could not be read gets no row and is not charged: it did not answer, it answers with an error page, or it shows an anti-bot check instead of its pages. Turn on includeUnreachable if you prefer one row per input website, with status blocked or unreachable and the reason in detail.
  • Inputs that are not public websites are skipped: IP addresses, internal names, text that is not a domain. They are listed under invalidInput in the SUMMARY record.
  • The page limit is a cap on pages read. maxPagesPerDomain is the most pages that count for one website. When a website has to be reopened in a browser, the pages already requested are requested again by the browser, and no new page of a website is read once it has used three minutes of browser time.
  • A restart does not start over. Finished websites are remembered and are neither read nor charged again, and a run you stop half-way keeps every row already delivered. Websites that were not finished are listed under notProcessed.

Every run writes a SUMMARY record to its key-value store with the count of distinct websites requested (requested) and rows delivered (delivered), how many websites were read over plain requests (readOverHttp) and how many needed a browser (readWithBrowser), and the list of domains under noContactsFound, blocked, unreachable, notProcessed and invalidInput. Every distinct input is in the dataset or in one of those lists.

Input

FieldDefaultWhat it does
domainsrequiredDomains or URLs
maxPagesPerDomain4Most pages read per website (1 to 50)
includeGenericEmailstruefalse leaves out shared mailboxes (info@, support@, sales@, noreply@ and similar)
detectTechStacktrueFill the techStack column
onlyWithContactsfalseRows only for websites with contact details
includeUnreachablefalseAlso add rows for websites that could not be read
maxConcurrency3Websites open in a browser at the same time (1 to 10), for the ones that need a browser

How each website is read

  1. The home page is fetched with a plain request. Most company websites publish their contact details in the HTML they serve, and this is fast.
  2. The contact, about, team, imprint and legal pages that the home page links to are read next. If the limit is not used up and the website is being read with plain requests, common paths are tried, chosen by the website's country ending: /contact and /about in general, /kontakt and /impressum for .de, /contacto and /nosotros for Latin American endings, and so on. Imprint pages matter: websites in Germany and Austria are required to publish contact details there.
  3. A real browser is started only for the websites that need one: pages built by scripts, websites that refuse plain requests, and websites whose plain HTML shows no contact details but loads scripts that may add them. This is slower, which is why it is the exception. Browser visits go through Apify Proxy.
  4. If the website still shows an anti-bot check, nothing is read from it: what is written on such a page belongs to the anti-bot service, not to the company.

Where each column comes from

  • Emails. Mail links, page text, structured data, and addresses hidden behind Cloudflare email protection, HTML entities or "name [at] company [dot] com" spelling. Sample addresses shown inside a form field or an interface mock-up are left out, and so are addresses that belong to the tools a website is built with and the authors named in the licence notices of bundled libraries.
  • Phones. Each candidate is checked against the numbering rules of its country and written in international format. The country comes from the number's own prefix, or from the website's structured data, declared language or country ending. When none of those tells, only numbers written the North American way ((415) 555-2671) or with a + prefix are accepted. A bare run of digits is taken only when a label such as "Phone" or "Tel" stands next to it, and numbers that follow words such as "Order" or "Ref" are skipped. A call link written without a country prefix is read in the website's country; when that country is not known and the digits are also a valid number somewhere else, the number is left out rather than guessed. On a page that lists offices abroad with call links written without the +, only the links that cannot be numbers of the website's country are kept. This is a filter, not a guarantee: phonesVerifiedByLink tells you the numbers come from call links or structured data, which is more reliable than page text but still depends on how the website wrote them.
  • Social profiles. The profile the website declares in its structured data wins. Only the profiles declared for the organisation itself count, not those of an author or founder nested inside it. Otherwise the account whose handle matches the website's name is taken, and failing that an account is kept only when it is the single one of that network and sits in the header, footer or navigation, or is linked more than once. Share buttons and personal LinkedIn profiles in page links are ignored. This avoids returning a customer's or an author's account as if it were the company's.
  • Addresses. Only what the website declares in its structured data. Websites that write their address as free text return an empty list.
  • Contact page. The shortest contact address among the pages read, without tracking parameters.
  • Technology. Signatures in the HTML served. A website that documents or writes about a tool can match its signature.

Which platforms and tools it detects

detectTechStack matches the HTML served against 24 signatures:

GroupDetected as
Ecommerce and CMSShopify, WordPress, Magento, Drupal, Ghost, Wix, Squarespace, Webflow
CRM, chat and supportHubSpot, Intercom, Zendesk
Analytics and taggingGoogle Analytics, Google Tag Manager
PaymentsStripe
Front-end frameworksReact, Next.js, Vue.js, Angular, jQuery, Bootstrap, Tailwind CSS
Hosting and CDNCloudflare, Vercel, Netlify

Anything outside those 24 is not reported. The column is a shortlist for targeting (every Shopify store in your list, every website already running Intercom), not a full inventory.

Enriching a CSV of domains

Export the domain column from your CRM or lead list, paste it in, and join the result back on domain. Turn on includeUnreachable (and leave onlyWithContacts off) to get one row per distinct valid website, including the ones that could not be read.

Two columns split a list without opening a single website:

  • emails empty and hasContactForm true: a company that routes everything through a form. Worth a form submission or a LinkedIn message, not an email.
  • emailDetails.type: role addresses reach a department, personal addresses reach a person.

What it does not do

  • It returns what websites publish. It does not guess unpublished addresses; for a specific person's address, use Email Finder.
  • It does not test whether an address accepts mail; chain Email Verifier over the output for that.
  • It does not get past anti-bot checks that a normal browser visit cannot pass. Those websites are reported as blocked, never filled with made-up data.
  • It reads a handful of pages per website, not the whole website. An address published only in a blog post or a PDF will not be found.

Changes in version 0.2

If you used an earlier version:

  • Phone numbers are now validated and written in international format (+1 888 926 2289, before +1 (888) 926-2289).
  • Websites that show an anti-bot check no longer produce a row unless includeUnreachable is on.
  • New settings: onlyWithContacts and includeUnreachable.
  • domain is now the host in lower case, and emails lists addresses on the company's own domain first instead of alphabetically.
  • Social profiles are chosen more strictly, so a row can lose a profile that an earlier version returned when it could not be tied to the company.
  • New columns: status, method, detail, emailDetails, hasContactForm, phonesVerifiedByLink.
  • Most websites are now read with plain requests and a browser is used only when needed, so runs are much faster.

FAQ

Can it process many websites at once? Yes. Paste the whole list; all rows land in one dataset.

How many emails will I get per website? Whatever each website publishes. Many companies publish one shared mailbox, some publish a team page with personal addresses, and some publish only a form. The status column and the SUMMARY record tell you which is which.

Can I exclude info@ and support@ addresses? Yes, set includeGenericEmails to false.

What does a 1,000-website run cost? The per-row rate on the Pricing tab times the number of rows, plus the run-start fee. Websites that could not be read produce no row by default, so the number of rows is at most 1,000.

Can I cap the cost? Yes. Set a maximum charge for the run in the Apify Console; the run stops adding rows when it is reached and lists the websites it did not get to under notProcessed.

Is scraping contact information legal? The Actor reads only what companies publish on their own public websites. How you use the data (GDPR, CAN-SPAM, recipient preferences) is your responsibility in your jurisdiction.

Can I schedule it? Yes, with an Apify Schedule, and send the output to Google Sheets, Make, n8n, Zapier or a webhook.