How we test

Email files look simple but hide a lot of variation: different clients encode headers, attachments and HTML bodies in slightly different ways. A tool that only works on one export isn't good enough, so every converter here is checked against a growing corpus of real, sanitised test files before it's marked "live".

The test corpus

The corpus currently includes files exported from:

• Microsoft Outlook
• Gmail
• Apple Mail
• Mozilla Thunderbird

It covers plain-text and HTML bodies, inline images, regular and nested attachments,winmail.dat (TNEF) attachments, non-Latin subject lines and bodies (Arabic, Hindi, Japanese), and a few deliberately large files to check size limits per browser.

The full, pseudonymised corpus is being prepared for publication on GitHub so anyone can reproduce the results. Each tool page lists exactly which clients and file variations it has been tested with, and the date it was last tested.

What "tested" means

For every file in the corpus, the tool's output is checked by hand: does the text match, are attachments present and openable, does the layout survive conversion, and does anything silently fail? Anything that doesn't hold up is listed under that tool's "Known limitations" section rather than hidden.