Explain a five-field cron expression field by field, preview the next run times in your own time zone, and catch the day-of-month and day-of-week trap.
At 09:00, Monday to Friday.
Standard five-field cron.
Calculating…
Run times are worked out in your browser.
Cron Expression Explainer runs entirely on your device — nothing is uploaded.
You’re on the Free plan. Compare plans
Cron syntax is five fields of terse punctuation that nobody remembers between the times they need it: minute, hour, day of month, month, day of week. This reads an expression back in words, one field at a time, so you can check that what you wrote is what you meant before a job runs at the wrong hour for a month.
There is one rule that catches almost everyone, and it is the reason this tool exists. When both the day-of-month and the day-of-week fields are restricted, cron runs the job when either matches, not when both do. So 0 0 1 * 1 does not mean "the first of the month, if it is a Monday" — it means "every first of the month, and also every Monday". A schedule meant to fire twelve times a year fires about sixty-four. The tool flags this combination explicitly whenever it appears.
It parses the five-field Vixie dialect, which is what crontab on Linux and macOS uses, and what Kubernetes CronJobs, GitHub Actions schedules, Vercel cron and most CI systems accept. Expressions with six or seven fields belong to Quartz and Spring, where a seconds field comes first; those are refused with an explanation rather than read as five fields, because silently dropping a field would describe a completely different schedule. The Quartz-only tokens L, W and # are refused for the same reason.
Next-run times are worked out in your own browser, from your own clock, and labelled with the time zone they are in. That matters more than it sounds: a prerendered "next run" would be the build server's idea of now, which is wrong for every reader. It also means these times are only the times your job will run if the machine running it shares your time zone — and around a daylight-saving transition, implementations genuinely disagree about a wall-clock time that occurs twice or not at all. The tool says so rather than inventing certainty.
Nothing is ever guessed. An impossible combination such as 30 February returns no run times at all rather than the nearest plausible date.
In order: minute (0–59), hour (0–23), day of month (1–31), month (1–12 or JAN–DEC) and day of week (0–6 or SUN–SAT, where both 0 and 7 mean Sunday). An asterisk means every value, and a comma, hyphen or slash builds lists, ranges and steps.
Almost certainly the day-of-month and day-of-week rule. When both fields are restricted cron fires when either matches, not both. Leave one of them as an asterisk unless you genuinely want the union of the two.
No, and it refuses them explicitly rather than reading the first five fields. Quartz and Spring put seconds first, so dropping a field would shift every other meaning and describe a different schedule. The L, W and # tokens are declined for the same reason.
Yours — they are computed in your browser from your own clock, and the zone is named above the list. The machine running your job has its own time zone, and unless it matches, these are not the times it will fire.
Implementations differ, which is why the tool says so instead of picking one. A daily job scheduled in the hour that is skipped in spring may not run at all, and one in the hour repeated in autumn may run twice. Scheduling jobs in UTC avoids the whole question.
Every fifteenth value of that field starting from its lowest — so */15 in the minute field means :00, :15, :30 and :45. Steps work in any field, and can be combined with a range: 0-30/10 in the minute field gives :00, :10, :20 and :30.
Because no date matches it. 0 0 30 2 * asks for 30 February, which never occurs, so the honest answer is an empty list. The tool searches five years ahead before concluding that a schedule never fires.