What counts as a number here
A number is a run of digits, optionally signed, optionally grouped, and optionally carrying a decimal
part: 42, -3.5, 1,234.56 and 1.234,56 are all numbers.
The only genuinely ambiguous part is the punctuation, because 1,234 means one thousand two
hundred and thirty-four in one convention and one-point-two-three-four in another. That is why the number
style is a setting rather than a guess — the same characters carry different values, and only you
know which your text uses.
Once matched, each number is parsed to its actual numeric value, which is what makes the sum and average possible. The list you get back is real numbers, not strings that happen to look like numbers.
How to extract numbers from text
- Paste the text. Drop any text into the box above — an invoice, a report, a log, a table copied from a page. The numbers do not need to be on their own lines or formatted in any particular way.
- Set the number style. Choose whether your numbers are written 1,234.56 (comma groups, dot decimal) or 1.234,56 (the European style). This is the one setting that matters, because the two styles read the same digits as different values.
- Read the totals. The count, sum, average, minimum and maximum appear above the list, worked out from every number found — which is usually the reason you pulled them out in the first place.
- Copy or export. Copy list puts the numbers on your clipboard. Save as CSV writes one number per row, ready to drop into a spreadsheet.
What comes out
- Every number in the text, in the order it appears, or sorted low to high if you ask.
- The count — how many numbers were found.
- The sum, added up from their real values.
- The average, minimum and maximum, so you have the range at a glance.
Two options change what you get: whole numbers or decimals only, when you want to separate counts from measurements, and keep original formatting, which returns the numbers exactly as written rather than normalised. The CSV export always writes the parsed values, so a spreadsheet reads them as numbers.
What is matched, and what is not
Matched: positive and negative numbers, with or without a decimal part, in either
grouping style. In 1,234.56 mode the comma is a thousands separator and the dot is the
decimal point; in 1.234,56 mode they swap; in plain mode there is no grouping and a single
dot is the decimal point.
Not matched as numbers: phone numbers, IP addresses and dates, which have their own
shape and their own tools — the phone number extractor, the
IP address extractor and pattern matching with the
regex extractor. A currency symbol or a percent sign next to a number is
not part of the number and is left out, so $99.99 yields 99.99.
Why the text is never uploaded
The text you paste is matched and the numbers are added up in your browser. No server takes part, so nothing is transmitted.
That matters when the numbers are not public — figures from an unpublished report, a client’s invoice, a payroll export. Adding them up should not mean sending them to a stranger’s server first, and here it does not.
What this tool does not do
Three boundaries:
- It does not do maths beyond the summary. Count, sum, average, minimum and maximum are provided; anything more belongs in a spreadsheet, which the CSV export feeds directly.
- It does not understand units or currency. It reads
10 kgand$10both as10; it cannot tell that one is a weight and the other a price. - It does not read numbers written as words. “Forty-two” is text, not a number, and is not converted.
Why people extract numbers from text
- Totalling figures pasted from a document or an email that is not already a spreadsheet.
- Checking an invoice or a report — pulling every figure out and summing it to confirm a stated total.
- Quick statistics — getting the average and range of a column of readings without setting up a spreadsheet.
- Pulling measurements out of prose — the numbers buried in a paragraph of description.
- Cleaning data — extracting the numeric values from a messy export before analysis.
The number extractor compared with the other text tools
This tool is for quantities. For values that are made of digits but are not really numbers, use the tool built for their shape: the phone number extractor for phone numbers, the IP address extractor for addresses, and the regex extractor for reference codes, order numbers and any pattern with its own rules. Those tools preserve the exact string; this one reads the value so it can add it up.
If your numbers are locked inside a PDF or a document, get the text out first with the PDF text extractor or the document text extractors, then paste it here.
Number format standards and separation ambiguities
The tool extracts integers and decimal numbers. Locale conventions differ: 1,234.56 (US/UK), 1.234,56 (continental Europe) and 1 234,56 (French) all represent the same value. The extractor normalises recognised formats to a decimal point in the output.
Three edge cases: (a) an eight-digit string like 20230101 is returned as a single number unless separated by a recognised delimiter — if it represents a date it must be split during post-processing; (b) a leading minus sign is captured when preceded by whitespace or the start of a line, so negative values are returned with their sign; (c) digits embedded in identifiers such as ID123, V2.3.1 or 0x1A2B are skipped when the digit run is bounded on both sides by non-separator word characters, to avoid fragments of version strings and product codes appearing in the output.
Frequently asked questions
How do I extract all the numbers from a block of text?
Paste the text into the box above and click Extract numbers. Every number is pulled out, and the count, sum, average, minimum and maximum are worked out for you.
Can it add up the numbers it finds?
Yes. The sum is shown automatically, along with the average and the range, because totalling the numbers is usually the reason for extracting them.
What is the difference between the two number styles?
1,234.56 uses a comma to group thousands and a dot for the decimal; 1.234,56 swaps them, as much of Europe does. The same digits mean different values, so choosing the right style is what makes the totals correct.
Does it handle negative numbers and decimals?
Yes. A leading minus sign is kept, and decimal parts are read in whichever style you selected. You can also filter to whole numbers or decimals only.
Will it pick up phone numbers or dates as numbers?
It reads any run of digits as a number, so a phone number or a year will be included. For those specifically, the phone number extractor and the regex extractor are the better fit.
Does it upload my text?
No. The numbers are matched and summed inside your browser, which is what makes it safe for figures from a private report or invoice.
Can it read numbers written as words?
No. It reads digits, not words, so 'forty-two' is treated as text and skipped. Only 42 is a number to it.