Four-timestamp calibration
Each sample records the browser’s send and receive readings plus the server’s receive and send timestamps.
Accuracy and trust
How Time.ms estimates accuracy
Time.ms shows exact current time by calibrating your browser against a network time reference, then displaying measured delay and estimated uncertainty.
Lowest-delay sample selection
Time.ms selects consistent samples with the lowest network delay and finds the time range they have in common.
Quantified error range
Time.ms publishes a numeric error range only when the browser timer and selected samples provide enough measurable information.
What the displayed time represents
After calibration, the page clock keeps running steadily, without following later changes to your device’s system clock.
How network delay affects precision
When available, the browser uses the actual request start and first response byte for calibration and round-trip delay (RTT). If detailed browser timing is unavailable, it uses the full request interval, giving a wider, more cautious bound. This round-trip delay is not directly divided by two to calculate the error range.
Why the device clock offset stays stable
Device clock offset is the difference between calibrated time and the device clock in the latest accepted calibration result. It updates only when a new result is accepted.
What happens when a source is unavailable
The browser checks the time server’s response. If a server is unavailable, it tries another with a new set of measurements; results from different servers are kept separate.
Privacy boundaries
Time synchronization does not require an account or precise device location. Browser geolocation is optional and only used after you click the device location control.
Displayed error range: the larger of the measurement estimate and the network-delay bound, rounded up to the next millisecond.
A measured high-resolution browser timer and a consistent shared time range produce an explicit error range. A low-resolution timer, inconsistent samples, or an initial calibration can still produce a clearly marked conservative estimate; only cases that cannot be quantified hide the number. The value stays fixed for the current calibration result and does not include device-clock drift accumulated afterward.
How to read the sync status
- Network latency
- The lowest round-trip latency among the selected valid measurements. It describes network conditions during the test, but it is not the time-error range.
- Time error
- The estimated time-error range, calculated from the overlap of the selected measurements’ time ranges and the measured browser and server timing limits.
- Offset
- The difference between the device clock and calibrated time in the latest accepted calibration result.
Check your own device
Run the clock drift test to measure your current offset. If the result is large, follow the device clock guide and test again.
Accuracy FAQ
Why does Time.ms show an accuracy estimate?
A browser cannot observe one-way network delay directly. Each four-timestamp sample therefore defines a possible time range. Time.ms compares several samples and uses their shared midpoint and half-width with measured timing limits. Repeated measurements show whether the result is consistent and whether more samples are needed.
Why can my device clock offset change?
The displayed offset updates only when Time.ms accepts a new calibration result.
Does Time.ms need precise device location?
No. Exact time synchronization does not require device geolocation.