AR packaging: the print is fixed. What it plays is not.
AR packaging is a printed box face, label, sleeve or hang tag that plays video, 3D, text and sound pinned to it when a shopper points a phone camera at it. With Interart it runs in the phone browser: the shopper scans a QR code printed on the pack, nothing is installed, and the panel’s own artwork is the tracking marker, so no marker is added to the design. Flat faces only — a bottle or a can is not this tool today.
Four things a box face can say that the print cannot.
The pack is already in the customer’s hands, at reading distance, with a phone in the other hand. That is the whole opportunity, and none of it needs a new print.
Ten seconds of the thing in use, standing on the front panel. A still photo on a box is a promise; a clip of it running is evidence. It can be re-cut after the run without touching the print.
The one fold, click or connection the exploded drawing never quite shows, played on the panel it is printed on. A 3D model of the part can stand beside the box and is seen from a new angle as the phone moves, for as long as the face stays in view.
Who made it, where, and why — a face and a voice for a brand that is otherwise a logo. Small producers get the most from this: the shelf cannot carry the story, and the box in someone’s hands can.
Where to get it again, a code for next time, the rest of the range beside it. Shown and spoken, not tapped: layers play rather than respond, so the address is read off the panel, and the shop’s own QR beside it does the tapping.
The same list for everything else you print: 10 ways to use AR on something you printed.
One flat, printed face is the marker.
The tracker matches what the camera sees against one flat image. Each surface below is that image. The ones that are not — bottles, cans, tubes — are further down.
The front panel of a carton is the strongest target on this page: it is drawn to be read across a shelf, so it usually carries the dense, contrasty artwork the tracker wants, and a panel is a rigid plane. One face is the marker — the one you photograph — and layers can stand off it, hang past its edges or sink behind it. A back panel that should play something of its own is a second design; the two can sit behind one entrance code on the pack, the way an exhibition puts a room of pieces behind one code at the door.
A flat label on a flat side works when it carries real detail. A jar’s side is a curve and a box’s is not. The tracking reference is 400 pixels on its longest side, so a label that is mostly white space with one small logo gives the tracker too little to hold; a busy one tracks.
A belly band, a flat sleeve or a swing tag is a small print, and small prints have a floor: detail finer than about three pixels at that scale is gone before tracking begins. A tag dense with illustration works the way a trading card does; a sparse one asks the shopper to stand closer than they will.
A bottle, a can, a tube or a pouch that does not lie flat is not one of these — why, and what “not today” means. The print spec in full: printing for AR. Several designs behind one code is the exhibition format: exhibitions.
Photograph the face, layer it, print the code on the pack.
-
1
Upload the panel artwork
The print file for the face, or a straight-on photo of the printed panel. The artwork itself is the marker — nothing is added to the design to make it trackable.
-
2
Layer what it plays
Video, audio, text and 3D, placed where they appear on the face and beyond its edges. Check it on a phone before anything goes to press.
-
3
Publish and print the QR on the pack
One link and one QR code per design. Print the code anywhere on the pack — the back panel, the tag, beside the face — and the camera still tracks the artwork. It opens in the phone browser already on the device — iOS Safari or Android Chrome.
-
4
Republish after the run
The print never changes and the layers live online. A new clip, a corrected step or a seasonal message replaces what every box already on a shelf plays, on the next scan. The code on the pack stays the same.
The same workflow, walked through on a poster: how to create an AR poster.
A QR to a web page, AR in your own app, or this.
Three honest ways to put something moving behind a pack. Which one is right depends on what the shopper has to be able to do, not on who is writing the page.
| QR to a web page | AR inside your own app | Interart | |
|---|---|---|---|
| What plays | A page, in a browser tab | Anything your developers build | Video, 3D, text and sound pinned to the printed face |
| What the shopper installs | Nothing | Your app, from a store | Nothing — it runs in the phone browser |
| Who builds it | Anyone with a web page | An SDK and a developer | You, in a visual editor |
| Taps, buttons, a quiz | Yes — it is a web page | Yes | No. Layers play; they do not respond |
| A curved bottle or can | Yes — only the code has to scan | Depends on the SDK | Not today — flat faces only |
| After the boxes are made | Edit the page | A new app version, and the store’s review queue | Republish; boxes on the shelf show the new version on the next scan |
| Cost of the trial run | Free | Developer time | Free: 3 published designs, 100 scans a month, commercial use included, and a small “Made with interart.ai” badge on the viewer that Pro removes |
What a browser can and cannot do, in depth: WebAR vs app AR.
Three things to settle with the printer and the plan.
Each of these has been the reason a piece did not work, and none of them can be fixed after the boxes are made.
A specular highlight replaces the artwork with a reflection of the room, and the tracker has nothing to match against a reflection. Gloss lamination turns a trackable matte panel into a mirror; spot UV and foil are fine over a border or a logo and bad over the detailed region. Matte or soft-touch lamination protects just as well.
The reference the tracker keeps is 400 pixels on its longest side. A busy hang tag works; a white tag with one small mark does not, and no setting recovers detail the camera never saw. Photograph the piece and test it on a phone before committing the run.
A scan is a phone opening the experience once; the same shopper back tomorrow is a second. The free plan covers 100 a month and puts a small “Made with interart.ai” badge on the viewer; Pro ($18/month) covers 10,000 with no badge, and Studio ($49/month) 100,000. Reaching the cap pauses the account’s published work for viewers until the 1st, and an upgrade lifts it at once. The cap is per account, so one busy product pauses everything the account has published — size the plan to the launch month, not the average one.
Above 100,000 scans a month it is quoted: say roughly how many units are going out and a person replies, usually within a day. The cap rule in full is on pricing; the finish and size rules are in printing for AR.
When AR packaging is the wrong tool.
Published because the rest of the page is only credible next to it. Each one is a real request, and each answer says what to use instead.
The tracker matches one flat image, and a label wrapped round a cylinder is not one: its edges bend away from the camera and its middle bulges toward it. Curved tracking is a lab plan with nothing in the engine yet — it covers a gently bowed card, and a full cylinder is outside even that — so the answer today is no. Put the AR on the bottle’s box or on a flat tag.
Layers play; they do not respond. A scene carries video, audio, images, shapes, text and 3D models, and none of them is a control, so a quiz on the box, a spin-to-win, a configurator or a shoppable tap needs a web page or an app. A plain QR to a web page does the tapping part better than any AR does.
A retailer’s app, a loyalty app, a scan-to-collect promotion inside something the shopper already has installed: that is an SDK and a developer, and none of this page applies. We publish a URL and a QR code, which is the right shape for a pack on a shelf and the wrong one for a feature inside an app.
Where each route wins, with dates: no-code AR · WebAR vs app AR.
Open a template, make it yours.
Interart’s AR templates are finished, working designs — open one in the browser editor, swap in your own artwork, and it’s yours. Each one teaches a technique with a built-in guided tour.
No account needed — sign up later and your work comes with you.
Put the demo on the box.
Start from a photo of the front panel — the first reveal takes minutes.
Questions, answered.
AR packaging is a printed box face, label, sleeve or hang tag that plays digital layers — video, 3D, text, sound — pinned to it when a shopper points a phone camera at it. With WebAR the experience runs in the phone browser: the shopper scans a QR code printed on the pack and points the phone at the printed face. No app, and no marker added to the design; the panel’s own artwork is what the camera tracks.
Not today. The tracker matches one flat printed image, and a label wrapped around a cylinder bends away from the camera at its edges. A bottle’s box, a flat hang tag or a label on a flat-sided container all work; the curved label itself does not. Curved-surface tracking is a lab plan with nothing in the engine yet, not a shipped feature.
No. Layers play; they do not respond to a tap. Video, sound, images, text and 3D can be shown and said, but none of the layers is a button, so a re-order address is read off the panel rather than tapped. The one tappable thing in a free-plan viewer is our own “Made with interart.ai” badge, which Pro removes. A plain QR to a web page does buttons better than AR does, and a game or a configurator needs an app.
High gloss stock, gloss lamination, and spot UV or foil over the artwork. Each puts a reflection of the room where the artwork should be, and the tracker cannot match a reflection. Matte or silk stock, matte or soft-touch lamination, and shiny treatments kept to borders and logos all track. Put the lamination type on the print order itself, because once the boxes are made it cannot be changed.
A scan is a phone opening the experience once, so count the scans you expect rather than the units you ship. The free plan covers 100 a month, with a “Made with interart.ai” badge on the viewer; Pro ($18/month) covers 10,000 with no badge, and Studio ($49/month) 100,000. Reaching the cap pauses the account’s published work for viewers until the 1st of the next month, and an upgrade lifts it at once. Above 100,000, ask for a quote before the run.
Yes. The print never changes — the layers live online. Republish with a new clip, a corrected assembly step or a seasonal message and every box already on a shelf plays the new version on the next scan. The QR code printed on the pack stays the same.