 ---   ---  |   | |   | ----- ----  -----
|   \ |   | |\  | |   | |     |   |   |  
|     |   | | \ | |   | |     |   |   |  
|     |   | |  \| |   | |---  |--/    |  
|     |   | |   | |   | |     | \     |  
|   / |   | |   |  \ /  |     |  \    |  
 ---   ---  |   |   /   ----- |   \   |

ASCII Converter

“ASCII converter” means two things: a picture turned into characters, or the code behind a character. This page is the code half — the art half is on the front page.

A character into its code

Direction

Your text 3 characters

Output Codes for your text

Decimal
72 105 33

Hex
48 69 21

Octal
110 151 41

Binary
01001000 01101001 00100001

Type or paste text and it prints each character’s code in all four bases at once; switch direction and it reads codes back into text. Switching direction moves the result into the box, so you can convert straight back and see that you get what you started with. It runs entirely in this tab — there is no request of any kind behind it.

Printable ASCII, 32 to 126

Ninety-five characters: the space, thirty-two punctuation marks, ten digits and both cases of the Latin alphabet. Decimal is what a specification quotes, hexadecimal is what a hex editor shows, octal survives in file permissions and in C escapes, and the binary is padded to eight bits so that a byte looks like a byte — ASCII itself only needs seven, and the leading zero is always a zero.

Char	Dec	Hex	Oct	Binary
space	32	20	40	00100000
!	33	21	41	00100001
"	34	22	42	00100010
#	35	23	43	00100011
$	36	24	44	00100100
%	37	25	45	00100101
&	38	26	46	00100110
'	39	27	47	00100111
(	40	28	50	00101000
)	41	29	51	00101001
*	42	2A	52	00101010
+	43	2B	53	00101011
,	44	2C	54	00101100
-	45	2D	55	00101101
.	46	2E	56	00101110
/	47	2F	57	00101111
0	48	30	60	00110000
1	49	31	61	00110001
2	50	32	62	00110010
3	51	33	63	00110011
4	52	34	64	00110100
5	53	35	65	00110101
6	54	36	66	00110110
7	55	37	67	00110111
8	56	38	70	00111000
9	57	39	71	00111001
:	58	3A	72	00111010
;	59	3B	73	00111011
<	60	3C	74	00111100
=	61	3D	75	00111101
>	62	3E	76	00111110
?	63	3F	77	00111111

Char	Dec	Hex	Oct	Binary
@	64	40	100	01000000
A	65	41	101	01000001
B	66	42	102	01000010
C	67	43	103	01000011
D	68	44	104	01000100
E	69	45	105	01000101
F	70	46	106	01000110
G	71	47	107	01000111
H	72	48	110	01001000
I	73	49	111	01001001
J	74	4A	112	01001010
K	75	4B	113	01001011
L	76	4C	114	01001100
M	77	4D	115	01001101
N	78	4E	116	01001110
O	79	4F	117	01001111
P	80	50	120	01010000
Q	81	51	121	01010001
R	82	52	122	01010010
S	83	53	123	01010011
T	84	54	124	01010100
U	85	55	125	01010101
V	86	56	126	01010110
W	87	57	127	01010111
X	88	58	130	01011000
Y	89	59	131	01011001
Z	90	5A	132	01011010
[	91	5B	133	01011011
\	92	5C	134	01011100
]	93	5D	135	01011101
^	94	5E	136	01011110
_	95	5F	137	01011111

Char	Dec	Hex	Oct	Binary
`	96	60	140	01100000
a	97	61	141	01100001
b	98	62	142	01100010
c	99	63	143	01100011
d	100	64	144	01100100
e	101	65	145	01100101
f	102	66	146	01100110
g	103	67	147	01100111
h	104	68	150	01101000
i	105	69	151	01101001
j	106	6A	152	01101010
k	107	6B	153	01101011
l	108	6C	154	01101100
m	109	6D	155	01101101
n	110	6E	156	01101110
o	111	6F	157	01101111
p	112	70	160	01110000
q	113	71	161	01110001
r	114	72	162	01110010
s	115	73	163	01110011
t	116	74	164	01110100
u	117	75	165	01110101
v	118	76	166	01110110
w	119	77	167	01110111
x	120	78	170	01111000
y	121	79	171	01111001
z	122	7A	172	01111010
{	123	7B	173	01111011
\|	124	7C	174	01111100
}	125	7D	175	01111101
~	126	7E	176	01111110

The codes below 32, and why 127 is DEL

Nothing under 32 prints. They were instructions to a teleprinter and several are still doing that job: 7 rings a bell, 8 backspaces, 9 tabs, 10 feeds a line, 13 returns the carriage, 0 is NUL, and 27 is escape — every ANSI color code this site can export begins with it. They have no glyphs, which is why they are not in the table.

127 is the odd one. It is a control code, DEL, but it sits at the top of the range instead of down with the others, and the reason is paper tape. 127 is 1111111: every hole in the row punched. A hole cannot be un-punched, so deleting a character already committed to the tape meant punching out every position it had, and readers were told to ignore that pattern. Delete had to be the all-ones code, which is why the printable range stops one short of the end.

How the converter works

The character-code converter walks the text one code point at a time and writes each one out in base ten, sixteen, eight and two. Line breaks are converted too — a line break is code 10, and the output is laid out one line per line of your text so that a multi-line paste stays readable. Going the other way, it splits on whitespace and commas, reads each token in the base you chose unless the token names its own, and refuses anything that is not a number in that base rather than guessing.

One honest caveat, and the converter states it on the page as well: above 127 what you get is the Unicode code point, not an ASCII code and not a byte. The first 128 code points of Unicode are the ASCII ones, and UTF-8 was built so that those 128 are still one byte each and still the same number — which is the whole reason ASCII text is already valid UTF-8. Above that the agreement stops: é is code point 233 and two bytes, and 😀 is code point 128512 and four.

A worked example

The three characters the code box starts with, Hi!: decimal 72 105 33, hexadecimal 48 69 21, octal 110 151 41, binary 01001000 01101001 00100001. That capital H is 72 and a lower-case h is 104 — a difference of 32, which is a single bit. Every letter’s lower-case code is its upper-case code with that bit set, which is why changing case used to be an OR rather than a table lookup.

Questions

What does an ASCII converter do?

Two different things, depending on who is asking. One kind turns a picture into ASCII art — characters arranged so that, from a distance, they make a shape; that one is on the front page, with six modes and no upload. The other turns a character into its ASCII code, the number the computer actually stores: 65 for A, 41 in hexadecimal. That is this page.

What is the ASCII code for a character?

Type it into the character-code box and you get the decimal, hexadecimal, octal and binary code at once. A is 65, a is 97, a space is 32, and the digit 0 is 48 — digits are not their own values, which is why subtracting 48 turns up in so much old C. The whole printable range is tabulated on this page.

How do I convert binary back to text?

Switch the character-code converter to “Codes to text”, choose Binary, and paste the numbers in, separated by spaces, commas or line breaks. 01001000 01101001 comes back as Hi. A token written 0x48, 0b1001000 or 0o110 is read in the base its prefix names whatever the buttons say, so a mixed list still works.

Why does my character have a code above 127?

Because it is not ASCII. ASCII is a seven-bit code with 128 positions, 0 to 127, and that is the whole of it: no accented letters, no curly quotes, no emoji. é is Unicode code point U+00E9, 233; 😀 is U+1F600, 128512. The converter reports the code point and says which characters are Unicode rather than quietly passing the number off as an ASCII code.

Is anything I type here sent anywhere?

No. The character-code converter is arithmetic on the string in the box: no request, no dependency, nothing stored. Once the page has loaded you can take the network away entirely and it still works. The image converter on the front page is the same promise for the same reason — your browser decodes the file and a worker in that tab converts it, because there is nowhere to send it to.

Which characters are printable ASCII?

Codes 32 to 126: the space, thirty-two punctuation marks, the ten digits and the two cases of the Latin alphabet — 95 characters. Everything below 32 is a control code that prints nothing, and 127 is DEL. The table on this page lists all 95 with their decimal, hexadecimal, octal and binary codes.

Where else to go

If what you came for was the other reading of the word — a picture turned into characters rather than a character turned into a number — it is on the front page, and everything about it is there too: six modes with real output under each, from plain shading through dithering, edge detection, outline, which draws one character per stroke instead of one per side of it, braille, which packs a two-by-four grid of dots into every character, and color as HTML, 24-bit ANSI or half-blocks. Nothing is uploaded there either.

Sizes are chosen in columns and the rows follow from them, which is the one part of that tool worth knowing before you open it: five presets from 24 columns to 240, a slider and a box to type an exact width into, and a readout giving the character count as well as the grid — the number that says whether a piece will fit a 2,000-character Discord message. What each size is for is set out there.

The rest of the site is short. About says who made this and why there are no accounts, credits lists the two fonts and their licenses, and privacy names in a table every event the site records — none of which can carry what you type into the box above, because nothing here leaves the tab.