Tiny Online Tools
What an MBOX actually is
An MBOX file is messages laid end to end, each introduced by a line beginning From — with no colon. There is no index, no length field and no table of contents. Thunderbird, Gmail Takeout, Apple Mail and most mailing-list archives all write this format, which is why an archive of ten thousand messages is a single file your mail client may take minutes to open.
Splitting it is mostly careful line reading. This tool finds every From line that is genuinely a separator — at the start of the file or after a blank line, and followed by something that looks like a header block — so a quoted From in the body of a message does not split it in half. Where a message carries a Content-Length header, that length is honoured instead, which is the only reliable way through a body that quotes a separator verbatim.
The dialect problem, stated plainly
Because a body line may itself begin From , mbox writers escape it as >From . Two dialects disagree about how far that goes. mboxrd also escapes lines already starting >From (as >>From ), so the escaping can be undone exactly. mboxo escapes only the unescaped case, and cannot be undone without guessing.
An archive containing >From but no >>From is consistent with both. This tool tells you which dialect it detected, says plainly when the evidence is ambiguous, and lets you choose the unescaping yourself. It is the sort of thing that silently corrupts quoted text if it is guessed at, so it is not guessed at here.
Naming the output
The filename template takes {index}, {date}, {time}, {from}, {subject} and {messageid}, and three names are previewed live as you type. Every token is sanitised on its own before it is substituted, so a subject containing slashes or a .. cannot compose a path; the finished name is capped, and duplicates get a numeric suffix. Every name ends .eml.
What is lost, honestly
The From_ line is not part of an RFC 5322 message — it belongs to the mbox container. It is dropped from every extracted file, and with it the envelope sender and the delivery timestamp it carried. Everything below it, headers and body alike, is written out byte for byte.
Privacy
The archive is read in your browser. No message leaves the page.