It was 1997. I was Chief Resident.
Standing in line in the cafeteria, a colleague stopped me.
“I have to go to Boston on the 14th of next month - so I need to have that night off.”
“umm .. ok”
I scribbled a note to myself on a napkin - hoping I didn’t lose it. I was making the call schedule on Excel - and I wondered why there wasn’t software that did this.
So I made software that did this.
And then - with the help of someone with better software development skills - iterated it for a few years, got some folks using it - and let it do its job. For over 20 years.
They say old software doesn’t rust and that’s kinda true. It worked. People used it .. it was working ..
And then 2025 happened, Claude Opus (and now Fable 5) got better .. and over the last few months, I’ve dug in once again (I have the time) at the far side of my career (now semi-retired) to add features, hone the user experience, and (finally) release a new version of the software 26 years after our initial release.
Now with a feature we dreamed about two decades ago: the schedule creates itself.
One click. It reads your group’s own history — who’s eligible, who’s off, how coverage actually works — and produces a complete, fair draft of next month. Every cell carries its reason: “balances weekend load,” “respects approved time-off.” Hand your group the fairness report and let the math speak.
Tell it your constraints in plain English — “Dr. Lee is out the 10th through the 14th; nobody works more than two weekends in a row” — and it listens.
And the part I care about most: the AI drafts. It never decides. A human approves every change, and nothing ships until you say so. I didn’t want to take clinicians out of the loop. I wanted to take the tedium out of it.
A 60-resident program or a four-provider practice. On-call or office hours or both.
The hand-built call schedule was an artifact of software that couldn’t do better. That excuse just expired.
If your chief resident or office manager still builds it by hand → oncalls.com. 30 day free trial.