A Thunderbird MBOX and a Gmail MBOX Are Not the Same File

Both are standard MBOX and both open in the same reader. What differs is everything around the format: Thunderbird gives you one file per folder, Gmail gives you one file with everything in it; Thunderbird organizes by folders, Gmail by labels stored in a header; and a Thunderbird file usually still contains mail its owner deleted years ago.

Knowing which one you are holding changes what you should expect to find.

One file each, or one file total

Thunderbird keeps one MBOX per folder: Inbox, Sent, Archives, each a separate extension-less file inside the profile, with a .msf index beside it and subfolders in a matching .sbd directory. To have “all the mail”, you need all the files.

Google Takeout gives you one file with everything selected in it — inbox, sent, spam and trash together — split into several parts only when it is very large. Nothing is missing, but nothing is separated either.

The practical consequence: if someone hands you “the Thunderbird archive” and it is one file, you probably have one folder of it. That is a common and expensive misunderstanding when the missing folder is Sent.

Folders versus labels

Thunderbird’s folders are exclusive: a message is in one, the way a piece of paper is in one drawer.

Gmail’s labels are not. A message can be in Inbox, starred and tagged Receipts at once, and Sent, Spam and Trash are themselves labels. Takeout encodes this in a non-standard X-Gmail-Labels header — a comma-separated list on each message — so each message appears once, carrying all its labels with it.

A reader that ignores that header shows you every message and loses the entire organization: twenty years of mail as one undifferentiated list. Mbox Viewer reads it and rebuilds the labels in the sidebar with counts, and label: works in search alongside from: and the rest.

The deleted mail that is still there

This one surprises people, and it matters legally as often as it matters technically.

Thunderbird does not remove a deleted message. It flags it in X-Mozilla-Status and moves on; the bytes stay until someone compacts the folder. An external reader has no reason to honour a Mozilla-specific flag, so it shows what is in the file — which can include mail the owner is certain they deleted in 2018.

Depending on the situation, that is either a recovery you are grateful for or something to handle carefully before the file goes anywhere. Either way, decide deliberately: compact first if those messages should not travel, or keep them knowingly. And count the messages afterwards, so you know what you actually have.

A Takeout export has no equivalent: it contains what was in the account at export time, including whatever was in Trash if you asked for it.

Counting them will not agree

Consequences of all of the above, when you try to reconcile two archives of “the same” mail:

• Thunderbird counts high if the folders were never compacted — deleted messages are still counted by a reader that walks the file.
• Gmail counts differently by design, because a conversation you think of as one thing is several messages, each with its own labels.
• Sent mail lives in different places. A folder in Thunderbird; a label in Gmail, mixed into the same file.

If you need one number, merge the sources into a single mailbox and deduplicate by Message-ID. That collapses the copies of a message that exists in both archives, which is the only way the arithmetic ever comes out.

Which one ages better

Both, honestly — they are the same open format, and that is the point of it. But they fail differently.

A Thunderbird profile is fragile in a specific way: the mail is in files whose names mean nothing outside Thunderbird, in a folder with a random name, and the .sbd structure is easy to break by tidying. Copy the whole Mail/Local Folders/ directory, not the files that look interesting.

A Takeout export is fragile in the other direction: it is a snapshot of a day, complete and self-contained, and immediately out of date. It is also the only one of the two that keeps working after the account is closed.

The habit worth having is the same for both: keep the original, name it so it means something in ten years, and open it once to check it is what you think it is.

---

Google and Google Takeout are trademarks of Google LLC.