Email extractor for a whole list at once

A list scraped out of a thread is never as long as it looks. The same person signs off five times, appears once in a mailto: link and once as visible text, and turns up again in a forwarded quote — so merging duplicates typically takes a third off the raw count before anything else happens. What the merge cannot see is the harder duplication: two genuinely different strings that arrive in the same inbox. Everything after a plus sign is a label the recipient invented, so sales+q3@acme.com and sales@acme.com are one mailbox at every large provider, and Gmail goes further and ignores dots entirely, which makes j.o.h.n@gmail.com and john@gmail.com the same person with no way to tell from the text. The tick box for +tags above handles the first case; the second is a Gmail house rule rather than anything in the address, so no tool can apply it safely to a domain that has not adopted it.

The domain table is there because the useful cut is almost never a single address, it is a whole category of them. Role accounts — info@, sales@, support@, postmaster@, abuse@, all named in RFC 2142 as mailboxes a domain is expected to run — belong to a job rather than a person, get read by whoever is on rota, and drag complaint rates up wherever they are mailed in bulk. A noreply@ address bounces by design. Consumer domains are the other obvious split: for anything B2B, the 27 providers behind the Drop personal mailboxes button are the addresses you keep for support and remove from a prospecting list. Once the shape of the list matters more than the addresses in it, the CSV export drops straight into the text to columns tool for splitting into named fields.

One honest caveat, because this page makes the mechanical part very easy and the legal part is not mechanical at all. Collecting addresses from a website and mailing them is treated differently almost everywhere: in the United States the CAN-SPAM Act names harvesting from web pages as an aggravating factor that raises the penalty for a violation, and in the EU and UK an email address that identifies a person is personal data, so marketing to it needs a lawful basis and usually consent gathered in advance. Extracting addresses from your own CRM export, a signed-up member list, or a document a client sent you is ordinary housekeeping; extracting them from strangers' websites to cold-mail is the thing the rules were written about. If the addresses are buried in running prose rather than listed — the name [at] company dot com shape — the extract emails from text page reads that form and checks each result instead of merely collecting it.