Skip to content
Lizely

Email Extractor

Extract, case-insensitively deduplicate, and optionally sort practical email addresses from pasted text locally.

Privacy: your files never leave your device. All processing happens locally in your browser.

How to use

  1. 1.Paste text containing possible email addresses and confirm it stays within the visible 500,000-character limit.
  2. 2.Choose first-appearance order or case-insensitive A–Z order, then select Extract emails.
  3. 3.Review the unique, duplicate, and rejected counts, then copy the newline-separated result.

About Email Extractor

Email Extractor finds common email-address shapes inside pasted text without uploading that text. Add contact notes, exported rows, logs, prose, or another text block, then choose whether results should stay in first-appearance order or be sorted A–Z. The output places one address on each line. Deduplication is ASCII case-insensitive, because addresses that differ only by letter case are rarely useful as separate extraction results, but the first spelling encountered is preserved. Thus [email protected] followed by [email protected] produces one result written as [email protected].

This is a practical extractor, not a complete RFC 5322 parser or an address-validity service. It recognizes an intentionally narrow ASCII form used by ordinary web addresses: letters, digits, dots, underscores, percent signs, plus signs, and hyphens in the local part; an at sign; and a DNS-style domain with at least two labels. It does not support quoted local parts, comments, escaped characters, domain literals such as an IP address in brackets, Unicode mailboxes, internationalized domain text, or punycode top-level domains. Some technically legal but unusual mailboxes are therefore excluded on purpose, and an extracted string is not proof that a mailbox exists or accepts mail.

Local-part checks prevent a dot at the beginning or end and reject consecutive dots. The local part is limited to 64 ASCII characters. Domain labels must contain letters, digits, or interior hyphens, must not begin or end with a hyphen, and must not be empty. The final label must contain 2–63 ASCII letters under this product's practical policy. A full candidate longer than 254 characters is rejected. These rules block frequent false positives such as [email protected], [email protected], [email protected], [email protected], single-label hosts, and one-letter final labels.

Common surrounding prose punctuation is not included. Commas, semicolons, closing brackets, quotes, exclamation marks, and similar delimiters naturally stop the candidate scan, while trailing sentence periods or ellipses are removed before validation. Malformed content is captured broadly enough to reject the whole candidate rather than quietly extracting a valid-looking suffix from it. Unicode text immediately attached to an address-shaped suffix is also rejected as one unsupported candidate instead of returning only its ASCII tail.

URLs are handled conservatively. A mailbox-shaped userinfo segment inside an HTTP, HTTPS, or FTP URL, such as https://[email protected]/private, is rejected rather than reported as a contact address. A mailto address remains extractable because its purpose is explicitly an email destination. The tool does not fetch links, inspect pages, follow redirects, query DNS, test SMTP servers, or send verification messages. It only analyzes the characters already present in the text area.

Limits are explicit. Input can contain at most 500,000 JavaScript characters. The text area does not silently stop accepting input at that point; instead, the visible counter marks an over-limit value and extraction returns an error without processing or truncating it. At most 1,000 unique addresses can be returned. If a text contains more, the operation fails with a message and does not expose a partial first thousand. Duplicate and rejected-candidate counts are shown after a successful run so the transformation is understandable.

Editing the input or changing sort order clears the prior list, error, and copy status. A failed extraction clears the old result before reporting the problem, so a list produced from earlier text cannot remain on screen as if it matched the current text. Copy all uses the browser clipboard asynchronously and only reports success for the current result; editing or leaving the page invalidates an in-flight copy status update. If clipboard permission is unavailable, the output remains selectable for manual copying.

Use this tool for cleaning ordinary contact lists, pulling addresses from notes, preparing a small import, or inspecting test data. Before sending email, review consent, privacy, suppression lists, jurisdictional requirements, and the source of every address. Extraction does not grant permission to contact someone, confirm ownership, identify role accounts, distinguish disposable mailboxes, or detect typographical domains. For high-stakes account creation or deliverability work, combine appropriate confirmation, DNS checks, bounce handling, and human review rather than treating a pattern match as verification.

Methodology & sources

A broad Unicode-aware candidate scan captures mailbox-like tokens, and boundary guards reject legal-looking fragments attached to a second at sign, invalid domain continuation, or combining mark. A documented practical ASCII validator then checks one at sign, a 64-character local-part limit, allowed common local characters, dot placement, total length, DNS-style labels, and a 2–63 ASCII-letter final label. Complete HTTP, HTTPS, and FTP URL ranges are collected in a separate linear pass so userinfo candidates remain excluded even in long URLs. Valid results use a lowercase deduplication key while preserving the first spelling; optional sorting compares those lowercase ASCII keys. Inputs above 500,000 characters or more than 1,000 unique results fail explicitly with no partial output.

Frequently asked questions

Does extraction prove an email address exists?
No. It only recognizes a documented practical character pattern. It does not query DNS, connect to a mail server, or send a verification message.
Why are some technically valid RFC addresses excluded?
The tool intentionally supports common ASCII addresses rather than every quoted, escaped, commented, literal, Unicode, or internationalized form allowed across email standards.
How does duplicate removal handle letter case?
Comparison uses a lowercase ASCII key, while the first spelling found is preserved in the output.
Are pasted addresses uploaded?
No. Candidate scanning, validation, deduplication, sorting, and clipboard preparation happen in the current browser tab.
What happens above the result limit?
More than 1,000 unique addresses causes an explicit error and no partial list is returned. Nothing is silently truncated.

Related tools

Text Tools guides

View all