Extract Numbers from Text

Extract numbers from text

Spreadsheets and invoices often bury the figures you need inside prose, logs or pasted reports. This tool scans any text and lists every number it contains — integers, decimals, negatives, thousands-grouped values and scientific notation — then cleans, sorts and totals them so you can drop the result straight into a spreadsheet.

How it works

The extractor matches numeric tokens including 42, 3.14, -42, 1,299.50 and 4.0e2. The matches then pass through your options: strip separators removes thousands commas so 1,299.50 becomes 1299.5; de-duplicate keeps each value once; sort numerically orders them ascending; and the separator toggle joins them with newlines or commas. Above the list it shows a running sum of every number found.

Example

Paste:

Invoice total 1,299.50, discount -50, tax 4.0e2.

With strip-separators on, the tool extracts:

Real-world scenarios where this saves time

Pulling amounts from a copied bank statement. Bank exports and email receipts often arrive as formatted prose rather than clean CSV. Paste the whole page, toggle strip-separators, and every transaction amount appears in a clean list ready to copy into a spreadsheet. The sum gives you an instant sanity check against the stated total.

Auditing a log file for response times or error codes. Server logs mix numeric values — milliseconds, HTTP status codes, byte counts — with hostnames and timestamps. Paste a section of the log, and the tool isolates only the numbers. Sorting them numerically surfaces the highest latency values or unusual status codes at a glance without writing a regex.

Reconciling a PDF invoice. When a PDF doesn’t export cleanly, copy-pasting the text layer produces a jumble of descriptions and prices. This tool pulls out every figure so you can verify the line items and confirm the total matches the grand sum.

Extracting measurement values from a research table. Academic papers sometimes contain tables pasted as plain text. The extractor strips the surrounding labels and gives you the numeric column to work with directly.

Edge cases worth knowing

• Dates and version numbers — something like 2024-06-15 or v3.1.4 produces the component numbers (2024, 6, 15 or 3, 1, 4), not the combined value. If your text contains dates, scan the output before summing.
• Currency symbols — a value like $1,299 strips cleanly to 1299 with separator-stripping on; the $ is a non-numeric character and is ignored.
• Scientific notation — 4.0e2 is converted to 400 for the sum. If you are working with very large or very small values from scientific documents, check that the exponent was parsed as intended.

Everything runs in your browser — nothing is uploaded.