History of Base64

The problem: email could not carry bytes

Electronic mail in the 1980s assumed the text of the oldest protocols: lines of printable characters, no bytes above 127, no raw binary. Attachments — images, executables, archives — simply could not travel. The fix had to speak the channel's language: rewrite every byte as printable text.

MIME gives the answer (1992–1996)

Nathaniel Borenstein's MIME standard — designed so email could carry rich content — defined a family of "content-transfer encodings" and chose 64 characters as the sweet spot: large enough to keep the overhead low, small enough to dodge every character that gateways mangled. The 1996 MIME specification (RFC 2045) fixed the exact alphabet, and Base64 attachments began flooding inboxes.

From email to everywhere

Once standardised, Base64 escaped the mailbox. Data URLs embed images in stylesheets; JWT tokens carry their JSON payloads as Base64 sections; databases, logs and APIs store binary blobs as text. RFC 4648 (2006) consolidated the family of base encodings and formalised the variants — including the URL-safe cousin that swaps + and / for - and _.

The lesson that stuck with it

Every generation rediscovers the same mistake: seeing Base64 output and calling it encryption. It is a courier uniform, not body armour — the encoding exists so data survives the journey, and it undoes itself just as easily on arrival. This site keeps it in the company of real ciphers precisely so the difference is impossible to miss.