Extract Numbers from Text
| Input | Caught | Why |
|---|---|---|
| 42 | 42 | plain integer |
| -3.14 | -3.14 | minus and decimal kept |
| 1,234.56 | 1234.56 | thousands comma read correctly |
| $100 | 100 | symbol stripped, number kept |
| v2.5 | 2.5 | context-blind quirk - see FAQ |
Paste a messy report, an invoice, a chat log or a column that lost its structure, and this tool pulls out every number: integers, decimals, negatives and comma-grouped thousands, listed one per line with a running count - and a live sum of everything found, which turns data cleanup into data arithmetic. Fifty line-item prices pasted in, one total out.
It is the third member of the extractor family: emails have their tool, web links have theirs, and numbers get this one, tuned for what digits need - thousand separators read correctly (1,234 is one number, not two), minus signs kept, currency symbols and units stripped off. Everything runs in this browser tab; nothing uploads, and your text stays on your own device.
How to use
- Paste your text - every number appears in the list live, one per line, with the count up top.
- Tick dedupe to fold repeated values; the sum line always totals whatever the list holds.
- Copy the block into a spreadsheet - one number per row pastes into a column directly.
Frequently asked questions
How do I extract numbers from text?
Paste the text above - numbers appear in the list live as you type. The matcher reads integers, decimals, negatives and comma-grouped thousands (1,234.56 comes out as one number, not three pieces), lists them one per line in order of appearance, and totals them at the end. Dedupe folds repeats when the same figure appears in every paragraph.
Does it really sum the numbers?
Yes - the sum line totals every number currently in the list, recomputed live as the text changes. That is the difference between cleanup and work: paste a column of invoice amounts and the total is already there, paste a stack of measurements and their sum sits under the list. The sum uses plain decimal addition on the extracted values in order of appearance.
Why does v2.5 come out as 2.5?
Because 2.5 inside a version string is a decimal number to the matcher, which sees digits, a dot and no context. The extractor is deliberately context-blind: it cannot know that v2.5 is a label while 2.5 liters is data. The practical fix is editing the text before or after - strip the version strings, or delete the stray entries from the list once copied.
What about currency symbols and units?
They stay in your text; only the number crosses into the list. $100 yields 100, 42 kg yields 42, -3.14 yields the negative. The matcher keeps the minus sign and decimal point and lets everything else - symbols, units, words - pass by, because the number is what a spreadsheet wants and the unit is what your column header already says.
Is my text uploaded anywhere?
No. The extraction runs entirely inside this browser tab - no server, no logging, no account. Your text stays in the page's local storage on your own device (up to 20,000 characters) so a refresh does not lose it, and nothing leaves your machine unless you copy it out yourself.