Planning around
changing capacity

Time is only part of the picture

Two free hours on a calendar can feel very different depending on your sleep and the work you've already done. ChronoZero starts there. It considers the effort a task calls for, when you might have room for it, and the recovery you'll need afterward.

Your calendar remains the record of your commitments. ChronoZero helps you work out how to use the time around them, with a plan you can change when it doesn't fit.

A plan that adjusts to your day

ChronoZero suggests when to take on demanding work and when to leave room for recovery. Its Power Zones describe different levels of effort in the plan. You can check in, correct its picture of your day and change a recommendation that doesn't fit.

Capacity is an estimate, with uncertainty. It isn't a direct measurement of your mind, and the model won't always get it right.

Your tasks and commitments

ChronoZero connects with Google Calendar and Google Tasks, Outlook Calendar and Microsoft To Do. The iOS app can also bring in Apple Calendar and Reminders. ICS and CalDAV offer other ways to connect. You choose the calendars and lists it may read.

The system uses this context to suggest suitable effort levels and task placement. You can review its interpretation and decide which items are fixed or flexible. Shared commitments stay in place, and changing a connected calendar requires your approval.

Turning notes into work you can plan

The Journal lets you write about your work in your own words. A note, project brief or learning goal can become a draft Path: a set of steps toward an outcome, with time estimates and prerequisites.

You can edit or discard the draft. It only enters planning after you accept it, and later steps wait until their prerequisites are complete.

Brünig: results from the current planner

Brünig is the planning model in the current alpha. We test it in controlled simulations, including comparisons run through the product's planning system. The results below describe those evaluations. They are not measurements of human productivity.

“Effective work” is a simulator-derived, capacity-weighted measure of work. It lets us compare schedules within an evaluation; its units are not hours. Results from separate evaluations should be read separately.

Compared with an earlier internal model

Brünig produced more effective work in all nine matched scenario pairs. Both models received identical inputs. One additional pair was excluded because the inputs did not match. The comparison model was used internally and was never deployed.

The median improvement was about 70%. The mean was roughly a doubling, pulled upward by two cases where the earlier model did almost no work. Nine pairs are a small sample, so these figures indicate a direction rather than a reliable effect size.

Median modeled capacity at the end of the recovery period was essentially unchanged in the matched comparison. That does not mean every scenario was free of trade-offs.

Compared with a fixed working day

A separate controlled benchmark compared Brünig with a rigid nine-to-five schedule, with sleep from 23:00 to 07:00 and a midday recovery break. Both used the same simulated population and evaluation horizon.

Effective work was about 26% higher, a difference of 2.83 units, with average sleep 0.22 hours lower. The control is deliberately inflexible; it does not represent every way people already organize their work.

Response to shorter sleep

In a controlled sleep-duration comparison, effective work increased from 10.31 after four hours of reported sleep to 19.01 after nine hours. Fixed calendar commitments were respected throughout.

After shorter sleep, the plan shifted toward steadier work and recovery. These are simulated outcomes, not a sleep prescription or a human trial.

Checks and known limitations

All ten scenarios in the alpha evaluation passed the checks establishing that the results were valid. Nine passed every behavioral check. All eight forced calendar blocks were honored, and all four paired tests of adaptation changed in the expected direction.

The tenth scenario failed our expectation for the response to a disruption. Its safety and sleep checks passed, but under our evaluation rules that still made the overall run a failure. Brünig was released after review of the individual results and the shortcomings of the earlier model.

There are other costs. One scenario showed more evening work under elevated modeled sleep pressure. Brünig also switched activities more often than the earlier model. The current evaluation does not cleanly distinguish useful transitions from unnecessary switching. These remain limitations of the current planner.

The limits of the alpha

We evaluate the planner in controlled simulations. Those results help us assess its behavior, but they don't establish that ChronoZero improves human productivity. Real days are less predictable, and people differ from the assumptions in any model.

ChronoZero remains an experimental alpha. Recommendations are open to review and override, and feedback from use matters. The system does not diagnose or measure medical conditions.

Readiness research has its own optional permission, off by default. Declining does not remove any product feature. You can withdraw through the web app's Readiness research setting. Research releases contain anonymous aggregates, not individual records; an already anonymous contribution cannot later be traced back and removed. The privacy notice explains the details.

Already approved? Sign in

Implementation details are temporarily withheld while we work on version 2.