Fast Phone Number Extractor

Extract phone numbers from any website in bulk. Give it one or more URLs and it returns every valid phone number found on each page, correctly parsed and formatted. It fetches pages through a managed browser-and-proxy backend, so you can choose a fast plain-HTTP mode or a full browser mode that renders JavaScript and bypasses anti-bot protection (Cloudflare and similar).

Use this actor to:

• Lead generation — collect contact numbers from a list of business websites
• Directory building — enrich a company list with phone numbers
• Data verification — check that a phone number listed elsewhere actually appears on a business's site
• Contact discovery — quickly find the phone number on any single page

Features

• Two fetch modes: fast request (cheap, no JavaScript) and browser (JavaScript rendering + anti-bot bypass)
• Multi-source extraction: tel:/sms: links, WhatsApp links, JSON-LD structured data, microdata, and page text
• Country-aware validation and national/international formatting via libphonenumber-js
• Per-page deduplication with the source of each number, original text, and match position
• Run statistics: finished/failed requests, pages with numbers, and unique-number events

How it works

1. The actor fetches each start URL in the selected mode.
2. It follows same-site links up to maxRequestsPerCrawl.
3. For every page it collects phone numbers from tel:/sms: links, WhatsApp links, JSON-LD, microdata, and visible text, then validates each with the selected country as a local-number hint.
4. It writes one dataset item per page and a STATS record for the run.

Request modes

Mode	JavaScript	Anti-bot bypass	Speed	Relative cost	Use when
request (default)	No	No	Fast	Low	The number is in the static HTML — most contact pages
browser	Yes	Yes	Slower	Higher	The page renders numbers with JavaScript or is protected by Cloudflare / a bot wall

Start with request. Switch to browser only for pages that return nothing in request mode or that block plain HTTP.

Getting more results

• Start from a contact or location page when you only need one business's number.
• Increase maxRequestsPerCrawl when numbers may live on deeper pages.
• Set the correct phoneCountryCode for local numbers without a + country prefix.
• Use browser mode for JavaScript-heavy or bot-protected sites.
• Results reflect phone numbers actually exposed by the website. Numbers embedded only in images are not read.

What data you get

Each page you crawl returns:

Field	Description
url	The page that was crawled (final URL after redirects)
title	The page's title
mode	The fetch mode used (request or browser)
httpStatus	The HTTP status reported for the page
phoneNumbers	Array of phone numbers found (see below)

Each entry in phoneNumbers includes:

Field	Description
number	Number in E.164 international format, e.g. +12125550198
nationalNumber	National significant number, e.g. 2125550198
formattedNumber	Number formatted for its country, e.g. (212) 555-0198
country	Detected country code, e.g. US
source	Where it was found: tel, whatsapp, json-ld, microdata, or text
text	The exact text or value the number was found in
position	Start/end offsets in the page text (only for text matches)

Input

Field	Type	Description
startUrls	array	Pages to extract phone numbers from
mode	string	Fetch mode: request (fast HTTP, no JavaScript) or browser (JavaScript + anti-bot bypass). Required, default request
maxRequestsPerCrawl	number	Maximum number of pages to visit, including links found on the start pages (default: 1)
phoneCountryCode	string	Country hint for parsing numbers that aren't already in international format, e.g. US, TR, GB (default: US)

Examples

One contact page (fast mode)

{
  "startUrls": [{ "url": "https://example-business.com/contact" }],
  "mode": "request",
  "maxRequestsPerCrawl": 1,
  "phoneCountryCode": "US"
}

JavaScript-rendered or bot-protected site (browser mode)

{
  "startUrls": [{ "url": "https://example.com/iletisim" }],
  "mode": "browser",
  "maxRequestsPerCrawl": 1,
  "phoneCountryCode": "TR"
}

Crawl several pages from multiple sites

{
  "startUrls": [
    { "url": "https://example.com" },
    { "url": "https://example.org/contact" }
  ],
  "mode": "request",
  "maxRequestsPerCrawl": 20,
  "phoneCountryCode": "GB"
}

Output sample

{
  "url": "https://example-business.com/contact",
  "title": "Contact Us",
  "mode": "request",
  "httpStatus": 200,
  "phoneNumbers": [
    {
      "number": "+12125550198",
      "nationalNumber": "2125550198",
      "formattedNumber": "(212) 555-0198",
      "country": "US",
      "source": "tel",
      "text": "(212) 555-0198"
    }
  ]
}

Notes

• phoneCountryCode only affects numbers written in a local (non-international) format. Numbers already written with a country code, e.g. +1 212 555 0198, are parsed correctly regardless.
• When the same number is found by more than one source, the highest-confidence source wins: tel/whatsapp links and structured data (json-ld, microdata) take precedence over free text.
• The default key-value store contains a STATS record with the mode used, finished/failed requests, pages with numbers, and unique-number events. uniqueNumbersFound is the sum of per-page unique numbers; the same number on two pages is counted twice.

License

ISC