Text to binary translator

Computers store numbers, not letters. A character encoding gives each character a number and says how to write it as bytes. ASCII, from the 1960s, covers 128 characters (English letters, digits, punctuation and control codes) in 7 bits, stored one per byte. A is 65, which is 01000001.

UTF-8 is the encoding the web and almost all modern software use. It keeps every ASCII character as the same single byte and uses 2 to 4 bytes for the rest of Unicode. The leading bits of the first byte say how many bytes the character has:

The waving hand emoji is code point U+1F44B. That is above U+FFFF, so UTF-8 uses four bytes. The code point’s 21 bits are spread over the x slots of the pattern:

U+1F44B         = 0 0001 1111 0100 0100 1011 (21 bits)
split 3/6/6/6     000  011111  010001  001011
into the pattern  11110000 10011111 10010001 10001011
hex               F0       9F       91       8B

The character-by-character table under the translator shows this split for every character you type. An é is two bytes (C3 A9); the Chinese character 中 is three (E4 B8 AD).

• Group into bytes. Each byte is 8 bits. Spaces between bytes make it unambiguous; without them the translator cuts the bits into groups of 8 and tells you if the length is not a multiple of 8.
• 7-bit groups. Some older examples drop the leading 0 of ASCII bytes (1001000 for H). With spaces between them they still decode here.
• Invalid bytes. Not every byte sequence is valid UTF-8; a lone 11111111 never is. The translator shows � for it and names the byte, which usually means a missing bit or text in another encoding.
• Hex dumps. Switch the byte format to Hex to decode 48 69, 4869 or \x48\x69.

// JavaScript (browser or Node): text → binary
[...new TextEncoder().encode('Hi')]
  .map(b => b.toString(2).padStart(8, '0'))
  .join(' ');                                   // '01001000 01101001'

// binary → text
new TextDecoder().decode(
  Uint8Array.from('01001000 01101001'.split(' '), b => parseInt(b, 2)));  // 'Hi'

# Python
' '.join(f'{b:08b}' for b in 'Hi'.encode())     # '01001000 01101001'
bytes(int(b, 2) for b in s.split()).decode()     # back to text

# Terminal
echo -n Hi | xxd -b      # 00000000: 01001000 01101001  Hi

charCodeAt() alone is not enough for this: it returns UTF-16 code units, so an emoji comes back as two numbers above 255 instead of its four UTF-8 bytes. Use TextEncoder to get real bytes.

The ASCII codes of common characters, which are also their UTF-8 bytes. For the full 0-127 table with control codes, see the ASCII table; to inspect any Unicode character, the Unicode converter. To turn numbers (not text) into binary, use the binary converter.