Python Unix Timestamp: time, datetime & Timezone-Aware Epochs

Get the current Unix timestamp

time.time() returns a float in seconds since the epoch. Because it is a float, a simple int(time.time()) gives whole seconds, and multiplying by 1000 before truncating gives milliseconds.

import time

sec = int(time.time())          # whole seconds, e.g. 1767225600
ms = int(time.time() * 1000)    # milliseconds
print(time.time())              # float with microseconds: 1767225600.123456

Convert a timestamp to a datetime

datetime.fromtimestamp() is the trap: without a tz argument it returns a naive local datetime — the same number means different wall-clock times on different machines. Always pass tz=timezone.utc for a deterministic result, or tz=ZoneInfo("Asia/Kolkata") for a specific zone.

from datetime import datetime, timezone
from zoneinfo import ZoneInfo

ts = 1767225600

naive_local = datetime.fromtimestamp(ts)                 # ⚠ depends on the machine's tz
aware_utc   = datetime.fromtimestamp(ts, tz=timezone.utc)   # 2026-01-01 00:00:00+00:00
aware_ist   = datetime.fromtimestamp(ts, tz=ZoneInfo("Asia/Kolkata"))  # 05:30+05:30

print(datetime.utcfromtimestamp(ts))  # DEPRECATED, and it returns a NAIVE object

Convert a datetime back to a timestamp

Only aware datetimes have a well-defined timestamp(). Calling it on a naive datetime makes Python assume local time — silently. If you have a naive object you know is UTC, attach the zone first with replace(tzinfo=timezone.utc).

from datetime import datetime, timezone

aware = datetime(2026, 1, 1, tzinfo=timezone.utc)
aware.timestamp()                     # 1767225600.0  ✓

naive = datetime(2026, 1, 1)          # no zone
naive.timestamp()                     # interpreted as LOCAL time — wrong result
naive.replace(tzinfo=timezone.utc).timestamp()  # 1767225600.0  ✓ explicit

Naive vs aware: the two kinds of datetime

• Naive — a wall-clock reading with no zone attached. It cannot be compared with aware datetimes (raises TypeError) and has no real instant.
• Aware — carries a tzinfo (UTC, a fixed offset, or a named ZoneInfo zone) and represents an actual instant.
• Rule of thumb: store and transmit UTC (aware), convert to a named zone only for display.

from datetime import datetime, timezone
from zoneinfo import ZoneInfo

now = datetime.now(timezone.utc)                    # aware UTC
ist = now.astimezone(ZoneInfo("Asia/Kolkata"))      # same instant, other clock
ist.utcoffset()                                     # datetime.timedelta(seconds=19800)

Native precision: microseconds

time.time() and datetime.timestamp() carry microsecond precision on CPython (the OS clock permitting). There is no native millisecond integer type — you derive it as int(time.time() * 1000), which is exactly what this site's converter detects from a 13-digit value.

Common pitfalls

• fromtimestamp without tz — naive local output; tests pass on your machine, fail on the server.
• timestamp() on naive objects — silently uses the local zone.
• utcfromtimestamp — deprecated in 3.12 and still returns naive; use fromtimestamp(ts, tz=timezone.utc).
• Mixing aware and naive — comparisons raise TypeError; normalize everything to aware UTC.

Try it live

Check any timestamp against these snippets in the epoch converter, or grab the current value from the Python snippet tab on the homepage.