Shelf Help

Appearance

Your first week with Shelf ​

This is for whoever just got a Shelf account and is looking at an empty workspace wondering where to start. It's not a feature tour — you'll find every feature eventually just by using the thing. It's an honest, ordered path through the first week, written from watching what actually works and what people trip over.

The short version: label and list the gear you'd notice missing, build one real booking, and stop there. Everything else can wait.

Day one: get in, and get ten things listed ​

Sign in, and if you're the first person from your company, you'll be offered the setup wizard. Do it — six short questions, maybe three minutes, and it seeds categories, locations and a first-run checklist that actually matches what you do rather than a generic template.

Then, before you touch anything else: add the ten things you'd hate to lose. Not your whole store. Ten. The camera body, the good tripod, the case of radios. This does three things at once:

  • It gets you through the assets-and-labels flow for real, on gear you actually care about.
  • It gives you something to print a label for and check scans properly before you commit to a hundred more.
  • It makes the workspace feel like yours instead of an empty product, which is the difference between finishing setup this week and never coming back to it.

If you already have a spreadsheet, don't retype it by hand — see Importing from a spreadsheet instead and come back here once it's in.

If you have a Brother or Zebra printer, connect it today too (Settings → Printers → Add relay). It's a five-minute job for whoever's comfortable running one command on an always-on machine, and every day it's not done is a day everyone else is stuck labelling gear by hand later. See Printing labels for the two traps worth knowing before you start (Documents/Desktop/Downloads folders, and giving the printer a fixed network address).

Days two and three: bring in the people, then the rest of the gear ​

Invite the humans first. From Settings → Team:

  • Anyone who needs their own login and role — an office manager, a warehouse lead — gets a proper invite.
  • Freelancers and casual crew who'll only ever hold gear, not log in, become non-registered members instead. No seat, no password, and they can still be handed custody of a case or named on a booking.

See Signing in for what each role can actually do — worth reading before you invite anyone, because Base and Self-service users don't see categories, tags, locations, reports or settings at all, by design.

Then keep adding gear, but resist the urge to get the whole warehouse in before you do anything else with it. A hundred assets with no bookings and no audits is just a very expensive spreadsheet. Add in batches, and use each batch as an excuse to print and scan, which is how you'll notice a labelling problem while it's still ten labels, not a thousand.

If you're adding several identical items at once — a dozen wireless mic receivers, the same model of light — build an asset model for it first and bulk-create from that rather than typing the same category and value in by hand a dozen times.

The first real job ​

Once there's real gear in Shelf, build a booking for an actual upcoming job — not a test one. Add gear by scanning rather than browsing the list; it's faster and it exercises the exact motion your team will use on load-in day. Check it out, take it to the job (or pretend to), check it back in. This is the moment Shelf stops being a database and starts being the thing your job runs on.

This is also where you'll meet custody for real, usually by accident — someone assigns custody of something that's already on a booking, gets the warning, and either backs out or means it. Read the warning. It's the one that costs a morning if you don't.

What can genuinely wait ​

  • A full historical backfill. The cable ties from 2019 do not need a Shelf record today. Add gear as you use it or as you have a spare hour, not as a weekend project.
  • A perfect category and tag taxonomy. Start with the wizard's seeded categories, rename what's obviously wrong, and let the rest settle over a few weeks of actual use. Renaming a category later doesn't break anything that's already filed under it.
  • Custom fields for everything you might one day want to prove. Add one when you actually need to answer a question — "when was this last tested and tagged" — not speculatively. Every custom field you add is one more thing on the asset form everyone has to skip past.
  • Audits. Run your first one once the gear you care about is actually labelled and located correctly. An audit on an unlabelled shelf just tells you what you already knew: none of it scans yet.
  • Reports. They read the data you've already put in — with three days of history there's nothing for them to say yet. Come back to Reports once you've got a month or two of bookings behind you.

What most people get wrong ​

Printing a hundred labels before doing the calibration print. It takes two minutes and it's the difference between a drawer of good labels and a drawer of labels that are all subtly the wrong size in a way that doesn't show up until they don't scan. Do it once per printer, first.

Running the print relay on a laptop that goes to sleep, or from inside Documents/Desktop/Downloads on a Mac. The relay has to be always-on and, on macOS, outside those three folders — see Printing labels for why. Get this right on day one and you never think about it again; get it wrong and every print job queues silently until someone notices.

Confusing custody with a booking, or a kit with a category. These are the mix-ups that make gear look lost when it isn't. Ten minutes with the glossary up front is cheaper than the confusion later.

Treating "missing" on an audit as an accusation. It isn't one — it just means nobody scanned that item before the audit was completed. Half the time it's out on a legitimate booking the audit's scope didn't account for. See Audits for how to work the missing list without anyone getting blamed for something that turns out to be fine.

One giant "Gear" category. It feels efficient on day one and is useless by week three, because every filter, every default label preset and every category-scoped custom field works against categories. Start with the wizard's seeded set and split further only when a category actually gets unwieldy — don't pre-optimise it before you've used it.

Everything on this site is written by Artefact Group for our own installation. Shelf is a fork of the AGPL-3.0 licensed Shelf.nu project — for upstream’s own documentation, see docs.shelf.nu.

Copyright © 2026 Artefact Group