What is a Unix timestamp and how is it calculated?

Short answer

A Unix timestamp represents time as an amount measured from one shared instant: 1970-01-01 00:00:00 UTC, known as the Unix epoch. The traditional unit is seconds, although many APIs and languages use milliseconds, so the unit is part of the data contract.

Three examples that establish the model

The timestamp does not store a weekday, a time zone, or a visual format. It stores a point on the timeline. Rendering that point as 01/01/2000, 1999-12-31 19:00, or 2000-01-01T00:00:00Z is a separate formatting and time zone decision.

How the value is calculated

Conceptually, subtract the UTC epoch from the UTC instant and express the difference in the chosen unit:

seconds timestamp = (UTC instant - UTC epoch) / 1 second
milliseconds      = (UTC instant - UTC epoch) / 1 millisecond

Exactly one day after the epoch is 24 × 60 × 60 = 86400 seconds. One day before it is -86400. A negative timestamp is not inherently invalid; it represents an instant before 1970.

What Unix time does not solve

• It does not identify the time zone used to enter a local date.
• It does not prevent missing or repeated local times during clock changes.
• It does not reveal whether an undocumented number uses seconds or milliseconds.
• It does not prove that the source clock was synchronized.
• Everyday Unix time normally does not count leap seconds.

Use the date and Unix timestamp converter to choose the unit and time basis explicitly. JWT NumericDate uses the same origin and counts seconds while ignoring leap seconds, as defined by RFC 7519.

Quick sanity check

A current value with about ten digits is probably seconds; thirteen digits often means milliseconds. That is only a clue. Confirm the producer’s contract before converting. Multiplying or dividing by 1000 must change the unit, not the instant.