Website Email & Phone Finder — checked before you send

Give it a company website. Get back every email address and phone number the company publishes, each one showing the page it was found on, plus a 0–100 send-risk score and a plain-English reason for every address.

Most contact scrapers hand you a list of strings and leave you to discover which ones bounce. This one tells you which inbox you are actually writing to — the enquiries desk, the recruitment desk, or an unattended no-reply — and every address is checked before you see it by the same engine that powers our Bulk Email Validator.

Checking is included, not an add-on. It is not a per-email extra, so your bill does not grow with how many contacts we find.

---

What you get on every row

column	what it is
email	the address, lowercased and de-duplicated
safe_to_send_score	0–100
recommended_action	send, send_low_volume, manual_review, suppress
recommendation_detail	one sentence explaining the decision
found_on	the exact page the address was found on
evidence	structured (schema.org), mailto (a link), or text — this also drives the score
primary_phone	best published number, in +44… form
primary_phone_national	the same number as a human would dial it
primary_phone_type	fixed_line, mobile, toll_free, voip, …
primary_phone_location	"London", "Berlin" — where the number is registered
all_phones	every number found
linkedin_url	the company's LinkedIn page, when it links to one
is_role_account / is_free_provider / is_disposable	the signals behind the score
pages_scanned	how many pages were read
status	ok, or a reason in plain words

Export from the Contacts view for a clean lead list, or the Phones view for one row per company.

How the score works

An address the company declared in its own structured data scores highest, a mailto: link it published scores nearly as high, and an address merely sitting in page text scores much lower — because that one might be a supplier's, a customer quotation, or years out of date. Where the address was found is treated as evidence, not decoration.

An info@ or hello@ on a company's own contact page is the address they ask strangers to use, so it is scored as the target rather than penalised as a shared inbox. A no-reply@ box reads 0 however prominently it is published.

Scores in the 60s and 70s are what a published, checked, non-probed address earns. Nothing reaches 90+, and that is deliberate: without mailbox-level SMTP verification we will not imply a certainty we do not have.

How it works

1. Reads the homepage, then picks up to four pages most likely to carry contact details — contact, about, team, imprint, get-in-touch — by looking at both the URL and the link text. Blogs, shops, carts and legal boilerplate are skipped.
2. Extracts addresses from schema.org structured data, mailto: links, and page text, undoing the obfuscation sites use: HTML entities,

   name [at] domain [dot] com
   , and Cloudflare's email protection. A scraper that ignores that last one returns nothing for pages that plainly show an address to a human.
3. Extracts phone numbers with Google's libphonenumber, so VAT identifiers, company registration numbers, order references and coordinates are not returned as phone numbers. The country is taken from the site itself.
4. Checks every address — syntax, domain, mail host (MX), disposable services, role inboxes, free providers — and scores it.

Which pages, and why so few

Homepage plus four is deliberate. A company that publishes no address on its homepage, contact, about, team or imprint page has not published one, and opening thirty pages to prove it would cost you money for nothing. Set Extra pages per website to 0 for a homepage-only scan, or up to 10 if you would rather dig.

What it does not do

• No mailbox-level SMTP verification. It cannot confirm that a specific inbox exists. It removes the provably bad addresses — invalid syntax, dead domains, disposable services, unattended no-reply boxes — and tells you honestly which ones it cannot be certain about. Use it as the first pass; send whatever survives to a per-email verifier if you need mailbox-level certainty.
• No guessing. It never invents firstname.lastname@company.com. Every address returned was actually published, and found_on shows where.
• No phone numbers a numbering plan does not recognise. A strict setting means the occasional real but oddly-formatted number is missed. That is the right trade: the permissive setting returns VAT numbers as phone numbers.
• No login-gated sources. Websites only.

Judgement, and how it differs

An info@ or hello@ address on a company's own contact page is the address they ask strangers to use, so it comes back send rather than flagged for review. A careers@ inbox is real and monitored but exists to receive job applications, so it is named as the wrong door. A no-reply@ box is suppressed — nobody reads it. A consumer mailbox (@gmail.com) published as a company's contact is usually a small business running on a personal inbox: deliverable, and labelled as the weaker signal it is.

Input

Paste domains or full URLs — acme.com, www.acme.com, https://acme.com/ all work and count as one website, charged once. Or point it at a CSV link, or the dataset of an upstream Actor (a Google Maps scrape, for example) and the website column is detected automatically.

Pricing

$0.002 per website — $2 per 1,000. Charged once per website, after its results are saved, whether it yields ten addresses or none.

• Duplicates are not charged twice. acme.com, www.acme.com and https://acme.com/ are one website and one charge.
• No per-email fee. Every address found is checked and scored at no extra cost, so a company with nine contacts costs the same as one with a single address.
• A website that publishes nothing still costs $0.002, because the work happened — up to nine pages fetched and read. You get a row saying so, with the contact-form URL where there is one.

Use it in a pipeline

Scrape companies with any Google Maps or directory Actor, feed that dataset in here, and every company comes back with contactable people attached.

You do not need to run the addresses through our Bulk Email Validator afterwards. It is the same engine, so it would repeat the checks you have already paid for and reach the same conclusions. That Actor is for a list you got somewhere else — a CRM export, a conference list, something a colleague sent you — where nothing is known about the addresses yet. For the creator side of the same job, there is YouTube Creator Finder.

Built by Salomon Labs. Every number traces back to where it came from, and the system says "unknown" rather than guessing.