A pomodoro timer that actually finishes on time. These spend their whole life in a hidden tab, which is exactly where a naively built timer breaks: browsers throttle background tabs to one tick a second and then roughly one a minute, so a timer counting down by ticks runs slow and interrupts you minutes late. This one records when the phase should end and checks the clock, so it is right however rarely the browser lets it redraw. The long break also lands after the right number of focus sessions — counting phases instead makes it arrive twice as often, which is the usual bug and takes a while to notice.
| Phase | Length | When |
|---|---|---|
| Focus | 25 min | Rounds 1–4 |
| Short break | 5 min | After rounds 1, 2, 3 |
| Long break | 15 min | After round 4 |
| All of them | Adjustable | The numbers are a starting point |
Yes, and that is the main reason to use it. Browsers throttle hidden tabs so aggressively that a tick-counting timer runs minutes slow — which is worse than useless for something whose whole job is to interrupt you at the right moment. This checks the clock rather than counting.
Twenty-five minutes of focused work, a five-minute break, and a longer break after four rounds. The numbers are a starting point rather than a rule — the value is in the enforced stopping, not in the specific figures, and all four are adjustable here.
After the configured number of completed focus sessions, four by default. Note that it counts focus sessions, not phases — timers that count phases give you a long break every other round, which is a common bug and quietly halves the structure.
If you allow it. Sound and browser notifications are both optional and off until you turn them on — permission is asked at the point of use rather than on page load, and a timer that makes noise without being asked is worse than one that stays quiet.
For the session. Nothing is stored on a server and nothing survives a reload, which is the trade for needing no account.
Try it, then adjust. Some work suits fifty-minute blocks; some people focus better in fifteen. The technique's value is the enforced break, not the number — which is why every duration here is editable.
A pomodoro timer spends most of its life in a background tab, which is exactly the condition under which a naively built one stops working. Browsers throttle hidden tabs: intervals are clamped to one a second, and in Chrome to roughly one a minute after five minutes. A timer that decrements a counter on each tick therefore runs slow in the background and finishes minutes late — precisely when you are relying on it to interrupt you. This one records when the phase should end and compares against the clock on each repaint, so the interval is only a redraw trigger and the remaining time is correct however rarely it fires. `performance.now()` is used rather than `Date.now()` because it is monotonic and cannot be moved backwards by a clock adjustment mid-session. The phase cycle has a specific off-by-one worth naming: the long break belongs after a number of completed *focus* sessions, not after that many phases. Counting phases makes it arrive twice as often as configured, which is the commonest bug in these and is not obvious until you notice you are getting a long break every other round. Only a completed focus session increments the round counter here; a break never does. The countdown display rounds up rather than down, so it reaches zero only when the phase is genuinely over — a timer showing 0:00 while still running for another two hundred milliseconds reads as broken. Notification and sound are offered rather than assumed, because a timer that makes noise without being asked is worse than one that does nothing, and browser notification permission is requested at the point of use rather than on load. Nothing is transmitted and nothing is stored on a server; settings live in the page for the session.