If you have ever pasted a string of 0s and 1s into a decoder and got garbage like 👋 back, this post explains why, and how to decode binary correctly by hand. I also built a free tool that does it for you: The Binary Code Translator.

The short version

To translate binary to text, split it into groups of eight digits. Turn each group into a number using place values (128, 64, 32, 16, 8, 4, 2, 1), then look up the character for that number:

01001000 = 64 + 8 = 72 = H 

That is all a binary decoder does. The details that break most tools are in what comes after plain ASCII.

A worked example, with a twist

Here is a short message:

01001000 01101001 00100000 11110000 10011111 10010001 10001011 

Paste it in and you get Hi 👋. That is seven bytes but only four characters. Here is why:

Bytes Value Character
01001000 72 H
01101001 105 i
00100000 32 (space)
11110000 10011111 10010001 10001011 U+1F44B 👋 (4 bytes)

The first three bytes start with a 0, which means "this byte is a whole character on its own."

The fourth is different. A byte starting with 11110 means the next three bytes belong to the same character, so a correct decoder reads all four together and gets code point U+1F44B, the waving hand. That rule comes from UTF-8 (RFC 3629). A decoder that reads every byte as its own letter turns one emoji into four characters of mojibake.

How many bytes per character?

Character Bytes Binary
A 1 01000001
é 2 11000011 10101001
😀 4 11110000 10011111 10011000 10000000

English letters take one byte, accented letters take two, most Chinese and Japanese characters take three, and emoji usually take four. The decoder reads the leading bits of each byte to know how many bytes to group, so you never need to mark where a character ends.

Why won't my binary translate?

Most "broken" binary is not broken. It is in a different format, or a digit got lost when someone copied it.

What you pasted What is going on
7 digits per group (1001001) 7-bit ASCII without the leading zero. A good decoder detects this.
5 digits per group (10101) A 5-bit code, not ASCII
A total that won't split evenly (12 digits) A digit is missing or extra. One lost digit shifts every byte after it.
A lone byte above 127 (10101010) In UTF-8 this is only ever the middle of a longer character, so on its own it is not a character
Groups of 3 or 4 digits Not 8-bit text. Usually octal or a puzzle cipher.

A quick sanity check: if you paste a 5-letter word and the decoder reports 11 characters, the code was probably split in the wrong place.

Where binary comes from

Binary is a way of writing numbers using only 0 and 1. Text only appears once you agree which number stands for which letter (A = 65, stored as 01000001). That agreement is the encoding, following ASCII and UTF-8. Leibniz described the binary number system in 1703, and in 1937 Claude Shannon showed in his MIT master's thesis that on/off switching circuits could do binary logic. Every computer since runs on that idea.

A handy pattern: every capital letter starts with 010 and every lowercase letter starts with 011.

The tool

The Binary Code Translator handles all of the above:

  • Paste binary with or without spaces and it decodes as you type
  • It detects 8-bit, 7-bit and UTF-8 and tells you which it found
  • Emoji, accents and other languages work
  • You can copy the result, download a .txt, share a decoded link, or open a .txt or .bin file up to 1 MB
  • It runs in your browser, so nothing you type is uploaded

Every example above was decoded in the tool and checked against RFC 20 (ASCII) and RFC 3629 (UTF-8).

Try it

Paste 01001000 01101001 and see what comes back.

Try it here: https://thecodetranslator.com/binary-code-translator/