Extract Dates

Paste any text and get every date and time it mentions, normalized to ISO 8601 and listed in order. Runs in your browser; nothing is uploaded.

Dates found

About this tool

Extract Dates scans a block of text, finds every date and time it mentions, and lists them normalized to ISO 8601 in the order they appear. Each result shows the original text, the normalized value, and whether it's a date, datetime, or time.

Everything runs locally in your browser via WebAssembly — your text is never uploaded.

Recognized formats

Invalid dates and times (e.g. Feb 30, 25:99) are skipped automatically.

A note on ambiguity

Purely numeric dates like 01/05/2024 are inherently ambiguous between month-first (US) and day-first (most of the world). This tool assumes month-first unless the first number can only be a day (> 12). For unambiguous results, prefer ISO YYYY-MM-DD.

FAQ

Can I make numeric dates parse day-first (European style)?

Not with a switch — the parser is fixed to month-first and only reads day-first when the first number is greater than 12 (so 25/12/2024 is correctly 25 December). If your text uses day-first dates with days ≤ 12, the safest fix is converting them to ISO YYYY-MM-DD before extracting.

How are two-digit years like 3/4/99 interpreted?

With a fixed pivot: 0069 become 20xx and 7099 become 19xx, so 3/4/99 normalizes to 1999-03-04 and 3/4/25 to 2025-03-04. Four-digit years are always taken literally.

Why is a date next to a time reported as one datetime?

When a date and time are adjacent in an ISO-style form (2024-01-05 14:30 or with a T), they're matched together and reported as a single datetime in YYYY-MM-DDTHH:MM:SS form. Standalone clock times — including 12-hour ones like 3pm or 9:05 AM — are reported separately as time, normalized to 24-hour HH:MM:SS.

What happens with impossible dates like Feb 30?

They're validated against the real calendar and silently skipped — Feb 30, 13/13/2024 and clock values like 25:99 never appear in the results, so you don't get false positives from version numbers or scores.

Developer & Automation Access

Run it from the terminal

Same engine as this page, headless — via the gizza CLI:

gizza tool extract-dates "We met on January 5, 2024 at 3pm, then again 2024-02-10 and 25/12/2024."

New to the CLI? Get gizza →

Open it by URL

Pre-fill and auto-run this tool with query parameters — the names match the API/CLI:

https://gizza.ai/tools/extract-dates/?text=We%20met%20on%20January%205%2C%202024%20at%203pm%2C%20then%20again%202024-02-10%20and%2025%2F12%2F2024.

Machine-readable descriptor: tool.json — title + parameters JSON Schema for agents.