Remove Numbers from Text
Strip numbers from text in your browser - remove every digit 0-9, or delete only whole-number words like 5 and 3.50, and collapse the leftover spaces.
- Free, no account
- No watermark
- No usage limit
About the Remove Numbers from Text
"Remove numbers" sounds like one job. It's really two, and the difference is exactly where cheap tools go wrong. Stripping every 0 through 9 in sight is the easy version, and it's what most free removers do. Trouble is, that same pass reaches inside item5, v2, route66, and your phone numbers and removes their digits along with the rest, handing you back item, v, route, and a lone -. If all you wanted gone was a loose page number floating in a paragraph, you just broke half your labels.
This one gives you both passes and lets you pick which one runs. Strip every digit when you genuinely want zero numerals left anywhere. Or strip only the standalone numbers, the words that are nothing but a number, and leave anything with a letter in it completely untouched. It all runs in your browser, so nothing you paste is uploaded, saved, or logged, and a live before/after count tells you how many digits came out, so you're never squinting at the result wondering whether it caught what you meant.
How to use
- Paste or type your text into the Input box. The cleaned version appears below and updates the instant you change anything, no "process" button to hunt for.
- Pick a mode. Remove all digits deletes every 0-9 anywhere, even glued inside a word. Remove whole-number words only deletes a chunk of text only when the whole chunk is a number, so
5and3.50go butitem5stays. - Decide about the spaces. Leave Collapse the leftover spaces on and the tool folds the doubled gap a removed number leaves behind and trims each line's edges. Turn it off when exact spacing matters, like an aligned table.
- Copy the result, or hit Clear to start over. The original stays in the input box until you clear it, so nothing is lost if you flip between modes to compare.
Two modes, and why word mode is the one you usually want
Take one line: Copy 3 of invoice 12 filed under ref AB48X.
Run remove all digits on it and you get Copy of invoice filed under ref ABX. The loose 3 and 12 are gone, which is what you were after, but the 48 buried inside the reference code AB48X went with them and left you ABX. If that code mattered, you just broke it and might not notice until later.
Run remove whole-number words only on the same line and you get Copy of invoice filed under ref AB48X. The freestanding 3 and 12 vanish, the reference code survives whole, and that's the pass you want most of the time when you're cleaning real prose, it leaves your version tags, product codes, and IDs exactly where they were. Most removers online simply don't offer this mode. They give you the sledgehammer and call it done, and you find out later that the digits are gone from your part numbers.
Word mode is also careful about the shapes real numbers actually take. It doesn't just hunt bare runs of digits, it counts the punctuation that belongs to a number as part of it. Decimals like 3.50 and 0.99 come out whole, so you never get a lonely . stranded in a sentence the way a plain digit strip leaves one behind. Negatives like -10 and signed values like +42 go as a single unit, sign and all. Even a thousands-separated 1,000.25 reads as one number and is removed cleanly, comma included.
A phone number like 555-1234 survives word mode, and that catches people out. It reads as a number to you, but the tool sees a string with a hyphen wedged in the middle, so it fails the whole-chunk-is-a-number test and stays put. Dates like 2024-01-15, fractions like 1/2, and IP addresses like 192.168.0.1 all stay for the same reason, none of them is a plain integer or decimal. Same goes for a number touching a comma, so a bare 3, at the end of a clause keeps its digit too. If you want the numerals out of any of those, that's digit mode's job. It strips the digits and leaves the separators standing, so 555-1234 comes back as -. So the rule that sorts it: word mode handles standalone whole numbers, while digit mode is what you want when the digits are tangled up in punctuation or codes.
The digits a lot of number removers skip
One quieter edge only bites if you work with multilingual or scanned text, but when it bites it's baffling until you know what's going on. The 0-9 you type are the Western digits, and they are not the only ones. Unicode carries dozens of other digit systems, Arabic-Indic (٠١٢), Devanagari (०१२), Bengali, Thai, and the fullwidth forms (０１２) common in Japanese and Chinese documents. They look like numbers because they are numbers, they just live at different code points than the plain ASCII ten.
A basic remover that matches only [0-9] walks straight past every one of those and leaves them sitting in your supposedly cleaned output, but this one strips them too. Digit mode removes every character Unicode files as a decimal digit, so a٠١٢b cleans down to ab and a fullwidth ４２ disappears exactly like a plain 42 would.
Frequently asked questions
Will it remove numbers spelled out as words, like "seven" or "twelve"?
No. It works on numeric characters, 7 and 12, not the English words for them. seven and twelve are ordinary letter-words, so both modes leave them exactly as they are. Turning spelled-out numbers into digits, or deleting them, is a separate job this tool doesn't try to do.
What happens to citation numbers in brackets, like [12]?
Word mode leaves [12] alone, because with the brackets attached it isn't a bare number. If you're cleaning academic text and want those reference markers gone, digit mode strips the 12 and leaves the empty brackets [], which you can clear in a quick second pass with a find-and-replace. Loose citation numbers with no brackets, like as shown in prior work 12, word mode removes directly.
The spacing looks off after a number came out. Why?
When a standalone number goes, the spaces that flanked it stay behind, so item 5 costs turns into item costs with a double gap. That is what the Collapse the leftover spaces option is for, folding the run down to a single space and trimming each line's edges. Turn it off and the gaps are kept byte for byte, which you want for a fixed-width table or aligned columns where character positions carry meaning. Line breaks and paragraph structure are preserved either way.
Is my text uploaded anywhere?
No. All of the work happens locally in your browser, so your text is never sent, saved, or logged. That makes it safe for private notes, an unreleased draft, or anything you simply can't hand to a third party. It even keeps working if your connection drops after the page has loaded, since nothing needs to leave your device.
Is there a limit on how much text I can paste?
No fixed cap. Because the processing runs on your own machine, you can paste anything from a single line to a long document without it being truncated or shipped off somewhere. A very large paste might take a brief moment since your device does the work, and if the tab ever feels sluggish, split the text into a couple of chunks and run them one after the other.
I want the numbers pulled out into a list, not deleted. Can it do that?
No, and it's worth knowing before you paste because it's the exact opposite operation. This tool throws the numbers away and keeps the words around them. Pulling the numbers out into a list of their own, and discarding the text, is what a number extractor does instead. Grab the right one for the direction you actually want, or you'll get the precise inverse of what you meant.
A number survived word mode when I expected it gone. What happened?
Look closely and it almost certainly has a letter, a hyphen, a slash, or a bracket touching it, v2, 555-1234, 1/2, [12]. Word mode only removes a chunk that is entirely a number, by design, so anything mixed in with other characters stays. Switch to digit mode if you want those buried digits gone as well.