Extract numbers from text — with exact totals
This number extractor pulls every number out of a block of text and lists it as a clean column you can paste into a spreadsheet. It reads numbers the way people actually write them: international thousands separators (1,234,567.89) and the Indian lakh–crore style (12,34,567.89), negative values with a hyphen or a true minus sign (−7.5), percentages, amounts in rupees, dollars, euros, pounds and yen, scientific notation such as 6.022e23, and ranges like 10–20.
Next to the list you get the count, sum, average, median, smallest and largest value, plus a separate total for each currency, so you can check an invoice, add up figures scattered through an email or audit a report without retyping anything. The totals use exact decimal arithmetic rather than floating point, so 0.1 + 0.2 comes out as 0.3, not 0.30000000000000004.
Everything happens in your browser. Bank statements, salary slips and sales figures are never uploaded.
How to use the Number Extractor
- Paste your text into “Text containing numbers”, or press “Open a file” to load a .txt, .csv, .md or .html file.
- Check “Number format”: keep the first option for 1,234.56 and Indian 12,34,567.89, or choose the comma-decimals option when your text writes 1.234,56.
- Use “Show” to keep only currency amounts, percentages, whole numbers, decimals or negative numbers, and pick how ranges such as 10–20 are read.
- Choose the order, plain values or numbers exactly as written, and tick “Unique values only” to drop repeats.
- Read the totals, then press “Copy numbers”, or download a CSV with each value, its type, currency and line number.
Number Extractor features
Indian and international grouping
Reads 1,234,567.89, 12,34,567.89 and 1,00,000 correctly, and shows the totals in international, Indian or no grouping.
Currency-aware
Recognises ₹, Rs., $, US$, €, £ and ¥ before or after an amount, and codes such as INR, USD, EUR and GBP, then totals each currency separately.
Words for big numbers
Optionally turns “₹4.5 crore”, “2 lakh”, “$1.2bn”, “3 million” and “5k” into full values such as 45000000 and 200000.
Skips what is not a quantity
Dates, times, version numbers, IP addresses, phone numbers and hyphenated codes are left out, with a count of what was skipped.
Other digit systems
Devanagari (१२३), Bengali, Arabic-Indic and full-width digits are read as ordinary numbers.
Exact totals
Sum, average and median are computed on the exact decimal values, so long columns of money add up to the last paisa or cent.
Number formats it understands
| In your text | Extracted value | Read as |
|---|---|---|
| ₹12,34,567.89 | 1234567.89 | Currency (INR), Indian grouping |
| $1,234,567.89 | 1234567.89 | Currency (USD), international grouping |
| 1,00,000 units | 100000 | Whole number (one lakh) |
| −7.5 or -7.5 | -7.5 | Negative decimal |
| 18.5% | 18.5 | Percentage |
| 15–20% | 15 and 20 | Percentage range — or the midpoint, 17.5 |
| 6.022e23 or 3×10^8 | 6.022e23 and 3e8 | Scientific notation |
| €1.234,56 (comma-decimals setting) | 1234.56 | Currency (EUR) |
| ₹4.5 crore (with “Apply lakh, crore…”) | 45000000 | Currency (INR) |
What it leaves out on purpose
A plain “find every digit” search turns the date 14/10/2026 into three numbers and a phone number into four. This tool recognises those patterns first, skips them, and tells you how many it skipped:
- Dates such as 14/10/2026 or 2026-10-14, and times such as 10:30 or 9:45 pm.
- Phone numbers like +91 98765 43210 or (415) 555-0132, and digit runs of ten or more digits. Tick “Include phone-number-like digit runs” to keep them as single values.
- Version numbers and IP addresses (v2.4.1, 192.168.1.1) and hyphenated codes such as ISBN 978-3-16-148410-0.
- Digits glued to letters in names and codes — A4, MP3, B2B, COVID-19. Units after a number still work: 5kg, 250ml, 2nd and 1920x1080 give 5, 250, 2, 1920 and 1080.
- Anything inside web links and email addresses.
Worked example
Paste “Rent ₹18,500, electricity ₹2,340.75 and internet ₹999; refund of -₹500 on 05/10/2026.” The list reads 18500, 2340.75, 999 and -500, and the date is skipped. The sum is 21,339.75, the average 5,334.9375 and the median 1,669.875 (halfway between 999 and 2,340.75). International and Indian grouping only differ from a lakh upwards: 125,000 versus 1,25,000.
Need more than a sum? Paste the list into the Average Calculator or the Standard Deviation Calculator, or reorder it with Sort Lines. To pull phone numbers and dates out of the same text, use the Text Extractor; for addresses and links there are the Email Extractor and URL Extractor.
Number Extractor FAQ
How do I extract numbers from text?
Paste the text into the number extractor. Every number is listed under “Extracted numbers” as you type, one per line by default, with the count, sum and average above it. Press “Copy numbers” and paste them into Excel or Google Sheets.
Does it understand Indian number formatting like 1,00,000?
Yes. Lakh and crore grouping (1,00,000 and 12,34,567.89) is read alongside international grouping, and the totals can be shown in Indian style. Tick “Apply lakh, crore, million, k…” to turn “₹2.5 crore” into 25000000.
Can I extract only prices or percentages?
Choose “Currency amounts” or “Percentages” under “Show”. Currency totals are kept separate for each currency, because adding rupees to dollars would be meaningless.
Why is a number missing from the list?
It was probably recognised as a date, time, phone number, version, IP address or code, or its digits were attached to letters, as in A4. The line under the totals says what was skipped, and phone-like numbers can be included with a tick box.
How do I handle numbers like 1.234,56 with a decimal comma?
Switch “Number format” to the comma-decimals option. Dots are then read as thousands separators and commas as decimal points. Plain values always use a dot (1234.56), so spreadsheets read them correctly.
Is my text uploaded anywhere?
No. The number extractor runs entirely in your browser, and nothing you paste or open is sent or stored.