Email Sanitizer

A list of addresses pasted from a spreadsheet, an export or a thread is almost never clean, and the defects are consistent enough to be worth naming, because each one causes a different failure downstream.

Surrounding characters come first. Addresses copied from a mail client arrive wrapped in angle brackets or preceded by a display name. Addresses at the end of a sentence carry a trailing full stop. Addresses from a CSV may carry quotation marks that were the file format rather than the data. Every one of these produces an address that is subtly invalid, and most mail systems reject the whole row rather than repairing it.

Whitespace and invisible characters are next, and they are the ones that resist diagnosis. A leading or trailing space is invisible. A non-breaking space inside an address, which happens when a list comes from a web page, makes it unequal to the same address typed by hand and invalid as an address. A list that looks correct and fails validation usually has one of these in it.

Case is a decision rather than a defect. The domain half is case-insensitive by specification. The local half technically is not, though in practice essentially every provider treats it as insensitive. Lowercasing the whole address is standard practice and is what makes deduplication work, since the same person written two ways should merge into one record.

Deduplication should come after cleaning, never before. Two entries that differ only by a trailing space or a capital letter are the same person and will not be recognised as such until both have been normalised, which is why a list deduplicated first still contains obvious duplicates.

A word on what cleaning cannot tell you: a syntactically valid address may not exist, and no amount of tidying reveals that. Verifying deliverability means asking the mail server, which is a different operation with its own etiquette and its own consequences for your sending reputation.

This runs in your browser. A list of email addresses is personal data about identifiable people, and it is exactly the category that should not be pasted into a third-party service to be tidied.