Time Doctor idle detection: how it decides you stopped working

What "idle" means to the agent

The agent watches for mouse and keyboard events. It does not care what you type or where you click — it registers only that input arrived. As long as something keeps arriving, the clock keeps running. When nothing arrives for longer than a set stretch, the agent stops trusting that you are at the desk and records that time as idle.

Two things follow from that. Idle is defined by silence, not by a low number — it is a hole rather than a small percentage. And because the agent reads nothing but input events, it cannot tell a genuine break from focused reading: to it, a coffee run and an hour spent studying a document on screen look identical, because both produce no events.

Where the idle threshold comes from

The length of silence that trips idle is not a fixed Time Doctor value — it is an admin setting your company chooses, commonly a few minutes. Some teams set it short and strict, others more generously. From the employee side you usually cannot see the exact number, so the safest assumption is that a couple of minutes with no input is enough to start the countdown.

Time Doctor also runs in more than one mode. An interactive mode shows you what is being tracked and prompts you when the timer trips; a silent mode tracks quietly in the background. Which one you are on is the admin's choice too, and it decides whether you ever see the popup at all.

• The threshold is an admin setting, commonly a few minutes.
• Interactive vs silent mode decides whether you see any prompt.
• You rarely see the exact idle value from the employee side.

The "are you still working?" popup

In interactive mode, when the idle timer trips, Time Doctor surfaces a prompt asking whether to keep tracking that time. It is the agent's way of resolving the ambiguity it cannot resolve on its own: it saw silence, and it is asking a human whether that silence was a break or just quiet work.

The prompt is a fork. Confirm while it is up and the minutes are kept as tracked time. Dismiss it, or let it sit unanswered, and the stretch is usually dropped or marked idle. In silent mode there is no prompt — the agent decides on its own, and the idle time simply does not count.

• Answered while it is up: the minutes stay.
• Dismissed or ignored: the stretch usually drops out.
• Silent mode: no prompt, the agent decides alone.

What answering late actually does

The detail that catches people is that the popup is retroactive — it is asking about a gap that has already happened. Answer quickly and the time is kept. But if you come back to a stack of dismissed or expired prompts, the minutes behind them are generally the ones at risk: they either fall out of your total or surface on the report as idle stretches.

This is why a long, genuinely worked stretch away from the keyboard — a phone call, a whiteboard session, reading a long document on paper — can quietly cost you time even though you were working the whole time. The agent did not see the work; it saw silence, and no one answered for it.

Idle is not the same as low activity

It is worth separating two things people tend to blur together. Low activity means the tracker saw input, just sparsely — you still get a full timesheet with a smaller percentage on it. Idle means the tracker saw nothing at all, and instead of a low number you get a hole where the time should be.

That distinction matters because the two read completely differently to a manager. A low percentage is a number to interpret; a gap is a missing piece. Idling as a general problem across trackers has its own write-up in the related links.

How a gap reads on your report

Gaps are more legible than averages. A steady 70% across the afternoon invites no questions, but a blank stretch in the middle of the day is a shape a reviewer's eye lands on immediately — and a blank stretch that repeats at the same time every day looks like a pattern rather than a one-off.

If a genuinely worked gap needs explaining, context is your evidence: an idle window that lines up with a scheduled call explains itself the moment someone can see the calendar next to it. Understanding how the idle rule works is what lets you talk about the gap on your own terms instead of apologising for it.

Where Reaction fits

Reaction sits in one narrow place: the stretches when you step away but the session should stay alive — the coffee, the stretch, the school run — where otherwise the idle timer would start and a popup you are not there to answer would quietly drop the time. It keeps a natural, human-like level of presence within a range you set, using varied cursor motion, a living keyboard rhythm, occasional scrolling and window switches, and it starts only when the computer is genuinely idle and steps aside the moment you touch the mouse.

What it does not do is make low-input work look like high-input work while you are actually at the desk — that is a limit of the metric, not something to promise around. What it does is keep an ordinary break from reading as an absence on your timesheet. Use it within your agreements; see the acceptable-use note below.