What does Base64 do, and when should you use it?

Some places only take text: the body of an email, a field in a settings file, a line of a web page's code. A photo or a PDF is made of bytes, and plenty of those bytes aren't letters or digits at all, so they can't ride along as they are. Base64 rewrites the bytes with 64 ordinary characters: capital letters, small letters, digits, plus and slash. That's where the name comes from.

It's like reading a phone number aloud instead of handing over the paper it's written on: the words go where the paper can't, and the listener writes down exactly the same number.

So use Base64 when the place you're sending data to asks for it. The receiver has to know the text is Base64, or it'll treat the letters as ordinary words.

Only the spelling. Base64 reads your bytes 3 at a time and writes each group as 4 characters. The official Base64 standard, RFC 4648, uses the word Man as its example.

Two things follow. The text is about a third longer than the data it carries, which matters when a field has a length limit; the output-size guide compares Base64 with other spellings.

And it isn't a secret. Anyone who sees Base64 can decode it, with no key or password to find, so a password written in Base64 is still readable by anyone. Encryption is what keeps data private; Base64 can carry encrypted bytes, but it doesn't encrypt them.

Base64 always writes whole groups of 4 characters. When your data runs out partway through a group, equals signs fill the empty spots. They aren't data; they tell the decoder where the bytes stop.

The standard uses padding, but some formats that ask for Base64 say to leave it off. If a value arrives without equals signs, follow that format's rules instead of adding or deleting them yourself.

Typed text has to become bytes first. An encoding called UTF-8 turns each character into one or more bytes, and Base64 spells those bytes.

A different text encoding would make different bytes, and so different Base64, from the same words. If your Base64 has to match someone else's exactly, learn how text encodings differ.

Plus and slash already mean something in web addresses and file paths. So a variant called Base64url swaps them for a hyphen and an underscore, and leaves everything else alone.

FluffyBench's Base encoder and its File to Base64 tool both write standard Base64, with plus, slash and padding.

To read Base64 back as words, swap the From and To choices. If the decoded bytes aren't valid UTF-8 text, converting to Text shows an error instead of garbled characters. Changing Base64 into another spelling, such as Base16, works for any bytes.

Every byte holds 8 bits, the 0s and 1s a computer stores. Base64 regroups them 6 at a time, because 6 bits make exactly 64 possible values, one for each symbol in its alphabet.