Skejuler builds the schedule for physician groups that run 24/7 coverage.

Someone has to build the schedule.
It shouldn't be you. Every cycle.

Skejuler learns how your group actually works — who's half‑time, who can't take Thursdays, who shouldn't be on service two weekends running. You explain it once, in plain English. It drafts every cycle from there. You check it and publish.

Talk to us

This week · ICU

Providers from a sample week
	Mon	Tue	Wed	Thu	Fri	Sat	Sun
R. Alderton	E	E	—	Call	L	—	—
M. Behrens	—	L	L	E	—	D	D
N. Okonkwo	L	—	E	E	Call	—	—
T. Reyes	—	Call	—	L	E	E	—

You train it once

Next cycle, it already knows.

The first time, you tell it the things that live in your head — the clinic days, the part‑time lines, the two people who can't cover together. You don't tell it again. Every cycle after that starts from everything it has learned, including the corrections you made to the last one.

What it already knows · cycle 4

Alderton — no Thursdays (clinic)

Nights spread evenly across the year

Behrens 0.5 FTE — half the load

Never two weekends back to back

Learned in cycle 1. Not re-entered since.

Say it in English

Say the rule. See exactly what it became.

Every rule you give it becomes a constraint you can read, edit, or switch off. Nothing about your schedule is settled in a conversation: the AI reads your English, and a deterministic solver places the shifts — either satisfying every hard rule, or telling you precisely which one it couldn't.

Hard rules are never broken. Soft ones it reports when it has to bend them.

“Alderton shouldn't do more than two nights in a row, and never a night before her clinic day.”

Compiled — 2 constraints

HARDmax 2 consecutive nights · Alderton

HARDno night before clinic day · Alderton

Editable. Inspectable. The solver enforces these — not a chat.

Talk to us

20 minutes. You'll talk to someone who can actually answer.

A real schedule

Coverage first — every slot, every day.

Early, late, call, weekends, across sites. It builds against the coverage you actually have to fill, and it leads with what it couldn't fill rather than burying it. Nothing publishes until the gaps are closed and you've said so.

Sample data, invented names. Real divisions run 24/7 coverage across multiple sites.

Mar 3 – Mar 7 · 5 of 12 providers shown

Sample week of shift assignments
	M	T	W	T	F
R. Alderton	E	E	—	Call	L
M. Behrens	—	L	L	E	—
N. Okonkwo	Call	—	E	E	E
T. Reyes	L	Call	—	—	Call
J. Whitfield	E	—	Call	L	L

Early · Late · Call — 24/7 covered, no gaps.

Here's how we computed this

We show the arithmetic. You decide what's right.

Load is measured against each provider's own share — hours weighted by FTE, weekends counted on their own clock. Every number opens up to the arithmetic behind it, and every one of them can be overridden. We don't tell you what's fair. That was always your call.

Fairness Ledger: seven providers' weighted hours drawn against their fair-share target, one part-time row expanded to show the arithmetic — 69 hours worked against an 80-hour target. — Every number here shows the arithmetic behind it. Sample data, invented names.

Where we actually are

A working product, taking a few groups.