How to Extract Email Addresses from Text, HTML and CSV Files

Extraction is a data-handling task, not a permission slip. Process content you own or have a legitimate reason to work with: your own inbox, a CRM export, event registrations, or contacts who gave you their address. If you plan to send marketing email, you also need consent that satisfies the rules where your recipients live, such as GDPR in Europe, CAN-SPAM in the United States, or the DPDP Act in India. Starting with that limit saves you from building a list you cannot legally use.

Every email address has the same basic shape: a local part, the @ symbol, and a domain with at least one dot, such as name@company.com. A good extractor scans for that pattern, ignores surrounding punctuation such as brackets, quotes, and trailing full stops, and returns each match on its own line. It should also cope with uppercase letters, plus-addressing (name+news@company.com), and subdomains (team@mail.company.co.uk).

What it should not do is invent addresses. A pattern match is a syntax check. It cannot tell you whether the mailbox exists, whether the person still works there, or whether they agreed to hear from you. Keep that distinction in mind when you look at a long output list. The count is a starting point, not a count of reachable people.

Plain text is the easiest case: emails, chat logs, meeting notes, or a block copied from a PDF. Copy the full text, including headers and signatures if they contain the addresses you need. Paste it into the Email Extractor. Processing happens in your browser, so the content is not uploaded.

1. Copy the full text, including signatures and forwarded headers when those hold the addresses you need.
2. Paste it into the extractor and wait for the list to appear.
3. Scan for obvious problems, such as an address glued to a word because a space was lost during copying.
4. Copy or download the result, then move on to cleanup before you import it anywhere.

Web page source and HTML email templates hide addresses in several places: visible text, mailto links, form attributes, and sometimes comments. Copying only the visible page text can miss the mailto links. Copying the raw source catches them.

1. Open the page or template source. In most browsers, right-click and choose View page source.
2. Copy the whole source and paste it into the extractor.
3. Expect noise. HTML often contains placeholders such as noreply@, example@example.com, or addresses that belong to the site platform.
4. Remove anything that is not a real contact before you merge this list with others.

The same address often appears once in a link and again in the visible text. HTML extraction almost always produces duplicates. That is normal, and it is easy to fix in the cleanup step below. Do not treat the raw line count as the number of people.

CSV files from CRMs, form tools, and event platforms usually have an email column, but it is rarely the only place an address appears. You might find two addresses in one cell separated by a semicolon, an address inside a notes column, or extra columns that were added later. Rather than fighting the spreadsheet layout, treat the file as text.

1. Open the CSV in a text editor, or select all cells in the spreadsheet and copy them.
2. Paste everything into the extractor. Commas, tabs, and semicolons are treated as separators, so each address comes out on its own line.
3. For several large exports, use a bulk extractor built for bigger batches, then run the same cleanup.
4. Keep the original file. Extraction gives you addresses. Your spreadsheet still holds names, companies, and the source of each row.

This approach also catches addresses hiding in columns you were not expecting. A secondary billing contact in a notes field is a common example, and it is often the person who can actually reply.

Raw extraction output is a starting point, not a finished list. Three short passes make a large difference, and they take less time than explaining a bad send later.

Remove duplicates

Combining several sources almost guarantees repeats, and the same person may appear in different capitalisation. Run the list through Remove Duplicate Emails so each address appears once. Duplicates are not only untidy. Our guide on why duplicate emails hurt campaign performance covers the effect on complaints and engagement.

Fix or drop malformed entries

Look for addresses with spaces, double dots, a missing top-level domain, or a trailing character that the pattern still accepted. A syntax check catches these quickly. Anything you cannot confidently correct should be removed rather than guessed at. Guessing a domain from a typo creates a stranger's address.

Filter what you do not need

Depending on your goal, you may want to drop role addresses such as info@ or admin@, separate company domains from personal webmail, or exclude disposable domains. The companion guide on how to segment business and personal email addresses walks through that split and how to treat each group.

Imagine you ran a webinar and have three sources: the registration CSV, a thread of attendees who emailed questions, and the HTML of a partner page listing speakers. You paste all three into the extractor and get 640 lines. After removing duplicates you have 512. A syntax check drops 9 malformed entries. Filtering out noreply@ and platform addresses removes another 6. You now have 497 real, unique contacts, and you still know which file each group came from.

The important part is not the final number. It is that webinar registrants who opted in and speakers listed on a partner page stayed separable. They are not the same audience, and they should not receive the same follow-up.

• Mixing sources that have different permissions. Keep opted-in registrants apart from people who were merely named on a page.
• Trusting the extractor blindly. A match tells you the string looks like an address. Verify before any large send.
• Throwing away context. If names or company fields matter, keep the original file next to the extracted list.
• Skipping cleanup. A list full of duplicates and typos is one of the faster ways to damage sender reputation.
• Rebuilding obfuscated addresses. If the author wrote "name at company dot com", they did not hand you a mailbox.

Contact data is personal data. Work on a trusted device, avoid pasting lists into tools that upload content to unknown servers, and delete working copies you no longer need. Browser-based tools that process data locally reduce that risk because the content never leaves your machine. When you are done, the original export still belongs in whatever system already governs access to it, not in a downloads folder you forget about.

Extracting addresses is quick once you treat every source as text: paste it, extract, then clean. The extract itself takes seconds. The value comes from the cleanup, which means removing duplicates, dropping broken entries, and keeping only the audience you actually have permission to contact. Do those steps the same way every time and each list you build will be smaller, cleaner, and more useful than the raw paste you started with.