How to use the digit converter
Step 1 — Set the source base first
Pick what your numbers already are from the From menu. The tool does not guess — 101 is a valid number in every base and means five, sixty-five or two hundred and fifty-seven depending on which one. Getting this wrong is the one mistake that produces a plausible answer rather than an error.
Step 2 — Choose one number or a list
One number gives you the full picture: the answer, the same value in every other base, and how it sits in memory. A list, one per line takes a whole column pasted from a spreadsheet, a log or a register dump and converts every line in one pass.
Step 3 — Pick the target base
Any of the eighteen bases can be source or target, so hexadecimal to binary works as directly as decimal to hex — there is no need to route through decimal yourself. The swap button reverses the two and carries the result across with it.
Step 4 — Read it as you type
There is no Convert button to press. The answer updates on every keystroke, and if a character does not belong to the source base the tool says which character and stops — it never guesses at what you meant.
Step 5 — Check any line marked Invalid
In list mode a line fails when it contains a digit the source base does not have — a 2 in a binary number, an 8 in octal, a G in hex. The message names the character. Nearly always it means the source base is set wrong, not that the number is bad.
Step 6 — Take the results away
Copy result, Download results as plain text, or Download CSV — which gives you every base for a single number, or an input-and-output pair for every line of a list.
About the digit converter
This converts numbers between eighteen numeral systems — binary, octal, decimal, hexadecimal, base 32 and base 36 among them — in any direction, one value or a whole list at a time. It is free, needs no account, and runs entirely in your browser.
Every answer is exact, at any size
Most converters run on ordinary JavaScript numbers, which hold whole values exactly only up to 9,007,199,254,740,991. Past that they start rounding, quietly. It is the sort of failure you never notice, because the answer still looks like a number.
This tool works in arbitrary-precision integers, so there is no ceiling at all:
9007199254740993in hexadecimal is 20000000000001. A converter that rounds gives 20000000000000 — the same digits, the final 1 lost.12345678901234567890is AB54A98CEB1F0AD2, not AB54A98CEB1F0800.- An 80-digit decimal converts to hexadecimal digit for digit, with nothing dropped.
If you are converting hashes, database identifiers, cryptographic values or anything longer than about sixteen digits, this is the difference between a correct answer and a convincing one.
Fractions, not just whole numbers
Put a point in and it is carried through properly. 0.5 in decimal is 0.1 in binary. FF.8 in hexadecimal is 255.5 in decimal. 3.14159 becomes 3.243F3E03 in hex.
The fractional part is held as an exact ratio while it is converted, so it never picks up rounding on the way through — you choose how many fraction digits to show, from none up to 32. Many fractions do not terminate in another base at all: a third is 0.333… in decimal and 0.0101… in binary, so the digit limit is where the answer stops, not where it becomes wrong.
How the number sits in memory
For whole numbers the tool also shows what programmers actually need: how many bits the value needs, how many bytes that is, and how it would be stored in an 8, 16, 32 or 64-bit slot — both as an unsigned value and as a signed one in two’s complement.
Two’s complement is how processors store negative whole numbers: the value is written as 2width minus its size, so the leading bit ends up carrying the sign. It is why −1 is 0xFF in a byte, 0xFFFF in sixteen bits and 0xFFFFFFFF in thirty-two, and why a signed byte runs from −128 to 127 rather than −127 to 128.
The unsigned and signed columns are shown separately because “does 255 fit in 8 bits?” has two correct answers: yes as an unsigned byte (0xFF), no as a signed one, where the largest value is 127.
The bases, and what the awkward ones mean
- Binary, octal, decimal, hexadecimal — the four everyone needs, plus every other base from 3 to 12, and 20, 24 and 26.
- Base 36 is the natural extension of hexadecimal: after the ten digits it keeps going through the whole alphabet, 0 to 9 then A to Z. It is the densest base you can write with unaccented Latin characters, which is why short URLs and licence keys use it.
- Base 32 comes in two forms here, and they are not interchangeable:
- Digits 0–V — the conventional positional base 32, following the same pattern as hexadecimal.
- RFC 4648 letters A–Z, 2–7 — the alphabet chosen for that standard because it omits the characters people confuse when reading aloud: no 0 against O, no 1 against I or L.
255 is
7Vin the first andH7in the second. Both are offered because earlier versions of this page used the RFC alphabet, and results people saved should still reproduce.
One important distinction. This tool treats base 32 as a positional numeral system — it converts the value of a number. That is not the same operation as RFC 4648 Base32 encoding, which turns arbitrary bytes into text and is what you meet in TOTP secrets and file hashes. For that, and for its more common relative, use Base64 Encoder / Decoder.
Converting a whole column at once
Paste a list, one value per line, and every line is converted independently in one pass. Results come back in the same order, so a converted column still lines up with the rows it came from, and blank lines stay blank so your spacing survives.
Lines that cannot be read are marked Invalid with the offending character named, and a count appears above the results. That is deliberate: a silent wrong answer in the middle of two hundred rows is far worse than a visible failure.
Where this fits with the other tools
For text rather than numbers, Base64 Encoder / Decoder and Text and ASCII Converter handle encoding. Roman Numeral Converter covers Roman numerals and Number to Words writes figures out in full — both of which this page once claimed to do and never did. For arithmetic in another base, Scientific Calculator has hexadecimal and binary functions.
Frequently asked questions
How do I convert binary to decimal?
Set From to binary, type the number, and the decimal value appears immediately — along with the same value in octal, hexadecimal, base 32 and base 36 in the table underneath.
How do I convert decimal to hexadecimal or binary?
Set From to decimal and To to whichever you need. Every other base is listed at the same time, so you rarely need to convert twice.
What number bases does it support?
Eighteen: every base from 2 to 12, plus 16, 20, 24, 26, both base 32 alphabets and base 36. Any of them can be the source or the target.
Why does it say my number is invalid?
Because it contains a character the source base does not have — a 2 or 9 in a binary number, an 8 in octal, a G in hexadecimal. The message names the character. Nine times out of ten the number is fine and the From base is set wrong.
Is there a limit on how large a number can be?
No. Values are handled as arbitrary-precision integers, so a forty or eighty-digit number converts exactly. This matters more than it sounds: converters built on ordinary JavaScript numbers go quietly wrong above 9,007,199,254,740,991.
Can it convert decimals?
Yes. 0.5 in decimal is 0.1 in binary, and FF.8 in hexadecimal is 255.5. Choose how many fraction digits you want, up to 32. Set it to “whole numbers only” and the tool warns you when it is discarding a fractional part rather than dropping it silently.
Does it handle negative numbers?
Yes, with a leading minus in any base. For the machine representation of a negative value, read the two’s complement column — that is what a processor actually stores.
What is two’s complement?
The standard way computers store signed whole numbers: a negative value is written as 2width minus its size, which makes the top bit act as the sign and lets the same addition circuitry work for both signs. −1 in a byte is 0xFF.
Which base 32 should I use?
Digits 0–V if you want conventional positional base 32. The RFC 4648 alphabet if you are matching output from a system that uses it, or reproducing a result this page gave you before August 2026. They give different text for the same number.
Can I convert a whole list at once?
Yes — switch to A list, one per line and paste a column. Order is preserved, blank lines stay blank, and any unreadable line is marked rather than guessed at. The CSV download pairs each input with its output.
Does it convert Roman numerals or write numbers in words?
No, and it never has — an earlier description of this page said otherwise, which was wrong. Use Roman Numeral Converter or Number to Words for those.
Is it free and private?
Free, with no sign-up and no limit. The conversion runs entirely in your browser, so nothing you paste is transmitted or stored.