Direct answer
ASCII (American Standard Code for Information Interchange, first published in 1963) gives a number to 128 characters, from 0 to 127. Codes 0 to 31 and 127 are invisible control codes (tab, newline, escape). Codes 32 to 126 are the printable characters: space, punctuation, digits and English letters.
You don’t need a website to convert between them. Every Linux and macOS terminal already has the tools:
printf 'Hi' | xxd -p # ASCII to hex → 4869
echo 4869 | xxd -r -p # hex to ASCII → Hi
printf '%d\n' "'A" # character to code → 65The printable ASCII table
Each row holds three codes exactly 32 apart. That is why A (65) and a (97) sit on the same row: uppercase and lowercase letters differ by a single bit, 0x20.
| Dec | Hex | Char | Dec | Hex | Char | Dec | Hex | Char |
|---|---|---|---|---|---|---|---|---|
| 32 | 0x20 | SP | 64 | 0x40 | @ | 96 | 0x60 | ` |
| 33 | 0x21 | ! | 65 | 0x41 | A | 97 | 0x61 | a |
| 34 | 0x22 | " | 66 | 0x42 | B | 98 | 0x62 | b |
| 35 | 0x23 | # | 67 | 0x43 | C | 99 | 0x63 | c |
| 36 | 0x24 | $ | 68 | 0x44 | D | 100 | 0x64 | d |
| 37 | 0x25 | % | 69 | 0x45 | E | 101 | 0x65 | e |
| 38 | 0x26 | & | 70 | 0x46 | F | 102 | 0x66 | f |
| 39 | 0x27 | ' | 71 | 0x47 | G | 103 | 0x67 | g |
| 40 | 0x28 | ( | 72 | 0x48 | H | 104 | 0x68 | h |
| 41 | 0x29 | ) | 73 | 0x49 | I | 105 | 0x69 | i |
| 42 | 0x2A | * | 74 | 0x4A | J | 106 | 0x6A | j |
| 43 | 0x2B | + | 75 | 0x4B | K | 107 | 0x6B | k |
| 44 | 0x2C | , | 76 | 0x4C | L | 108 | 0x6C | l |
| 45 | 0x2D | - | 77 | 0x4D | M | 109 | 0x6D | m |
| 46 | 0x2E | . | 78 | 0x4E | N | 110 | 0x6E | n |
| 47 | 0x2F | / | 79 | 0x4F | O | 111 | 0x6F | o |
| 48 | 0x30 | 0 | 80 | 0x50 | P | 112 | 0x70 | p |
| 49 | 0x31 | 1 | 81 | 0x51 | Q | 113 | 0x71 | q |
| 50 | 0x32 | 2 | 82 | 0x52 | R | 114 | 0x72 | r |
| 51 | 0x33 | 3 | 83 | 0x53 | S | 115 | 0x73 | s |
| 52 | 0x34 | 4 | 84 | 0x54 | T | 116 | 0x74 | t |
| 53 | 0x35 | 5 | 85 | 0x55 | U | 117 | 0x75 | u |
| 54 | 0x36 | 6 | 86 | 0x56 | V | 118 | 0x76 | v |
| 55 | 0x37 | 7 | 87 | 0x57 | W | 119 | 0x77 | w |
| 56 | 0x38 | 8 | 88 | 0x58 | X | 120 | 0x78 | x |
| 57 | 0x39 | 9 | 89 | 0x59 | Y | 121 | 0x79 | y |
| 58 | 0x3A | : | 90 | 0x5A | Z | 122 | 0x7A | z |
| 59 | 0x3B | ; | 91 | 0x5B | [ | 123 | 0x7B | { |
| 60 | 0x3C | < | 92 | 0x5C | \ | 124 | 0x7C | | |
| 61 | 0x3D | = | 93 | 0x5D | ] | 125 | 0x7D | } |
| 62 | 0x3E | > | 94 | 0x5E | ^ | 126 | 0x7E | ~ |
| 63 | 0x3F | ? | 95 | 0x5F | _ | 127 | 0x7F | DEL |
SP is the space character. DEL (127) is a control code, not a printable character.
The control codes you will actually meet
| Name | Dec | Hex | Escape | What it does |
|---|---|---|---|---|
| NUL | 0 | 0x00 | \0 |
End of a C string; separator for find -print0 and xargs -0 |
| TAB | 9 | 0x09 | \t |
Horizontal tab |
| LF | 10 | 0x0A | \n |
Line feed, the Unix newline |
| CR | 13 | 0x0D | \r |
Carriage return; Windows ends lines with CR LF |
| ESC | 27 | 0x1B | \e |
Starts terminal colour and cursor sequences |
| DEL | 127 | 0x7F | Delete |
You can see them with xxd. A tab, a carriage return and a newline:
printf 'a\tb\r\n' | xxd
# 00000000: 6109 620d 0a a.b..Three patterns worth remembering
- Digits:
'0'is0x30,'9'is0x39. Subtract0x30and you have the number. - Uppercase:
AtoZis0x41to0x5A. - Lowercase:
atozis0x61to0x7A, the uppercase value plus0x20.
Here is the letter A in binary. Bit 5 (the 0x20 bit) is the only difference between A and a:
printf 'A' | xxd -b
# 00000000: 01000001 AHex to ASCII and ASCII to hex in the terminal
# ASCII to hex: full dump, plain hex, or space-separated bytes
printf 'Hi' | xxd # 00000000: 4869 Hi
printf 'Hi' | xxd -p # 4869
printf 'Hi' | od -An -tx1 # 48 69
# Hex to ASCII
echo 4869 | xxd -r -p # Hi
printf '\x48\x69\n' # Hi
# Character to code, and code to character
printf '%d\n' "'A" # 65
printf "\\$(printf '%03o' 97)\n" # a
python3 -c "print(ord('A'), chr(97))" # 65 aOn Linux, man ascii prints the complete table, control codes included, in octal, decimal and hex.
Base64 is not ASCII, but it lives next door
Base64 turns any bytes into a safe subset of printable ASCII (letters, digits, +, / and =), so binary data can travel through email, JSON or URLs:
printf 'Hi' | base64 # SGk=
echo SGk= | base64 -d # HiWhere ASCII ends and UTF-8 begins
ASCII stops at 127, so it fits in 7 bits. UTF-8 was designed so that every ASCII byte means exactly the same thing in UTF-8. Everything else (accents, Persian, emoji) uses two to four bytes, all of them 0x80 or higher:
printf 'سلام' | xxd
# 00000000: d8b3 d984 d8a7 d985 ........Four Persian letters, eight bytes. xxd shows dots because those bytes are not printable ASCII.
Some bytes, such as 0xFE, can never appear in valid UTF-8. A Linux filename can still contain them, because to the kernel a filename is just bytes. When a tool “repairs” such a name, it swaps the byte for EF BF BD (the � replacement character), and the next tool in the pipeline can no longer find the file:
printf 'test_\xFEinvalid' | iconv -f utf-8 -t utf-8
# iconv: illegal input sequence at position 5I fixed exactly this bug in fd. The story, with the before and after hex dumps, is here: Your Filename Can Be Weird. Your Tools Shouldn’t Make It Weirder.
FAQ
What does ASCII stand for?
American Standard Code for Information Interchange. It was first published in 1963 by the American Standards Association, now ANSI.
How many bits does ASCII use?
7 bits, which gives 128 codes (0 to 127). In practice each character is stored in an 8-bit byte with the top bit set to 0.
What is the ASCII value of A and a?
A is 65 (0x41) and a is 97 (0x61). The difference is always 32 (0x20).
Is ASCII the same as UTF-8?
No, but UTF-8 is a superset of it. Any valid ASCII text is also valid UTF-8 with the same bytes. UTF-8 adds multi-byte sequences for every other character.
What is “extended ASCII”?
Codes 128 to 255 are not part of ASCII. Older code pages such as Latin-1 and Windows-1252 each used them differently, which is why the same byte can look different on different systems. UTF-8 replaced most of them.