What Extract Numbers helps you do

List unique integers and decimals found in the input, including numbers with grouped thousands. It pulls the numeric values out of mixed text so you can work with them as data. The values are returned in the form they appeared, so a grouped thousand keeps its comma.

Number extraction runs in your browser and does not upload your text to Haivix.

How it works

These tools scan the pasted text in the browser with a pattern or parser and return the matches in first-seen order. Extraction is a candidate list, not a verification step: an email-shaped string may not be deliverable, a URL may not resolve, and a number may be part of a date or an identifier. Nothing is uploaded, and the source text is never changed. The result is a snapshot of the pasted text; editing the source afterwards does not update the list.

How to use Extract Numbers

Paste the text containing numbers.
Extract the distinct numeric values.
Check ambiguous matches such as dates or IDs.
Copy the list for a spreadsheet or script.

When to use it

Pull measurements or quantities out of a document.
Collect numeric values from a log for analysis.
Build a unique list of IDs or codes from text.
Extract prices from a copied page for a quick check.
Gather figures from a report before charting them.
Check which numbers appear in a configuration snippet.
Check the numeric values in a copied invoice.

When not to use it

Do not use it as a validation step for amounts or ranges.
Do not expect dates, times or version strings to be parsed as single values.
Do not use it on numeric data where the grouping separator matters for import.

Example

A sentence mentioning 1,000 and 250.5 returns those two values as numbers. A date such as 2026-01-02 may produce several number-like fragments, and a phone number can look like a single large integer. Review the list before treating it as a data column.

What you can do with Extract Numbers

Find signed integers and decimals.
Recognize grouped thousands such as 1,000.
Return each value once, in first-seen order.
Keep the original formatting of each match.
Copy the list for a spreadsheet or script.
Run the extraction locally with no upload.

Example: check the matches on a sample

Paste a short sample first and count the expected matches by eye. Compare that count with the result before running a large file, because a missing match is easier to spot in a small sample.

Read a few matches in context to catch false positives. A phone-like number inside an order ID, or a URL inside a code snippet, is common and should not be treated as a contact or a live link.

Copy the list and keep the original. If the extraction feeds a mail merge or a report, verify a sample of the destinations rather than assuming every match is valid.

If the result is empty, check the input format before assuming the tool is broken. A bare domain without http, an address split across a line, or JSON with a trailing comma are common reasons for no matches.

For a list that will be contacted or published, keep the source and the extracted list side by side. The source is the only way to resolve a questionable match or explain why one was missed.

What to know before you start

Grouped thousands may be returned with the separator, and phone-like strings can be ambiguous.
It returns text that looks numeric and does not validate ranges or units.
It does not normalize the values or convert units.
It does not keep the surrounding label or column.

Common mistakes

Assuming every match is a meaningful quantity.
Feeding grouped thousands into a parser that does not accept commas.
Expecting dates or version numbers to stay intact.
Ignoring negative signs attached to a value.

Troubleshooting

If a number is split, it may contain an unusual separator or a space.

If a date produced several values, extract it with a date-specific tool instead.

If duplicates remain, the values differ by sign or formatting.

If the list is empty, the text may not contain a numeric run.

Extract Numbers