Convert MBOX to PDF — one PDF per email

An MBOX is a single file of messages laid end to end, each starting on a line that begins "From ". Mailward does not read that file as one string — browsers cannot hold a string much past 512 MB, which is where most viewers fail on Takeout exports — but scans it in 4 MB chunks once, remembers where every message begins and ends, and parses a message only when it is opened or exported. A PDF export therefore reads one message from disk, renders it, adds it to the ZIP and moves on; memory holds the message being written, not the mailbox.

Google Takeout adds labels. Every message in a Gmail export carries an X-Gmail-Labels header, and Mailward turns each label into a folder — including labels written in Japanese or Chinese — with the mailbox itself as the root. That structure is what the batch export follows: a label's folder becomes a directory of PDFs in the ZIP, sibling names that collide case-insensitively get a -2, and a message filed under two labels is written into both, since that is where a reader expects to find it.

Inside each PDF, the message reads as it would in the viewer: subject, From, To, Cc and the original Date line, the text — the plain part, or its HTML converted with paragraphs, lists, table rows and pictures kept when there is no plain part or the HTML shows pictures — and its attachments embedded. Japanese, Chinese and Korean messages, common in Gmail exports, print in their own scripts, the extra font fetched from this site only when a message needs it.

What it costs: listing, searching and reading the mailbox is free up to its first 500 MB, and three messages a day can be saved as PDF without paying. Exporting a folder, a label or the whole mailbox as PDFs is part of the one-time Pro license.