Binary ⇄ Text Converter

A character is a number. Unicode assigns every character a code point — the letter H is U+0048, decimal 72 — and a character encoding is the rule that turns that code point into bytes. A byte, also called an octet, is eight bits, and a bit is one digit in base 2. So the letter H under ASCII is the single byte 01001000, and under UTF-8 it is the same byte, because UTF-8 was designed to agree with ASCII over the first 128 code points.

There is no such thing as “the” binary for a character until you say which encoding you mean. The character é is one code point, U+00E9, and it is one byte in Latin-1 and two bytes — C3 A9 — in UTF-8. This page always states the encoding it used, and shows the code point, the byte and the eight bits of every byte rather than presenting a lookup you have to trust.

ASCII defines 128 code points, U+0000 to U+007F, in four blocks of 32: control codes 0–31, punctuation and digits 32–63, upper-case letters and a few symbols 64–95, and lower-case letters 96–127. Seven bits are enough for all of them, which is why ASCII is a 7-bit code.

The two facts exam questions actually ask for: A = 65 = 01000001 and a = 97 = 01100001. The difference is exactly bit 5 — decimal 32, 0x20 — so case conversion is a single XOR: 65 ⊕ 32 = 97.

The eighth bit of a byte was originally free, and the commonest thing to do with it was to make it a parity bit. That history is on the parity and checksum page: 7-bit ASCII plus one parity bit is a byte.

Every printable ASCII character, from space (32) to tilde (126), with its decimal value, its hexadecimal value and its eight bits. This table is in the page’s HTML rather than drawn by script, so it prints and it is readable without JavaScript.

UTF-8 encodes a code point in one to four octets according to its magnitude. The lead octet says how many octets follow, and every continuation octet begins 10.

• Surrogates. U+D800 to U+DFFF exist only as a UTF-16 mechanism and are not encodable.
• Code points above U+10FFFF. The four-octet form can express more than Unicode defines; the excess is invalid.
• Overlong forms. A code point must use the shortest template that fits it.

RFC 3629 is explicit about why the third rule matters: “a naive implementation may decode the overlong UTF-8 sequence C0 80 into the character U+0000 … Implementations of the decoding algorithm above MUST protect against decoding invalid sequences.” The WHATWG Encoding Standard then fixes what a browser does instead: emit U+FFFD, and “no other behavior is permitted”. This page’s decoder follows that rule by default and offers a strict mode that reports the offset where the byte stream breaks.

Four different numbers describe the same string, and confusing them is the source of most encoding bugs.

• Bytes are written most significant bit first.
• Groups of eight are separated by a space for reading; the space is not part of the data and a copied bit string has none.
• Hexadecimal is upper case.
• Code points are written U+ followed by at least four upper-case hexadecimal digits.

• RFC 3629 — UTF-8, a transformation format of ISO 10646 (opens in a new tab): the octet templates and the prohibition on overlong forms.
• WHATWG Encoding Standard (opens in a new tab): the decoder state machine and the U+FFFD replacement rule this page implements.
• The Unicode Standard (opens in a new tab), including its best practice for the use of U+FFFD.