WebAR vs app AR: what actually separates them.
App AR runs inside something a viewer installs; WebAR runs in the browser already on their phone, opened by a link or a QR code. The app can reach hardware the browser cannot — depth sensors, persistent world anchors, face meshes. The browser reaches people who will never install anything, which on a printed piece is nearly all of them.
Where each one genuinely wins.
Written to be usable by someone choosing, which means the left column has to win some rows. It does.
| WebAR (browser) | App AR (native) | |
|---|---|---|
| What the viewer does first | Scans a code, or taps a link | Finds the app, installs it, opens it, finds the experience |
| Realistic reach on a printed piece | Anyone with a phone | The share of people who will install an app to look at a poster |
| Image tracking on artwork | Yes — the print is the marker | Yes |
| Depth, persistent world anchors, face meshes | No — the browser is not given that hardware access | Yes, via ARKit and ARCore |
| Heavy 3D, physics, game logic | Limited by the browser and the download | Where it belongs |
| Changing it after the print run | Republish; the same code shows the new version | New build, new review, and only for people who update |
| Cost to ship one campaign | A page and a QR code | Development, two store listings, review, and maintenance |
| What happens when the vendor leaves | The link dies unless someone keeps serving it | The app is delisted and takes the experiences with it |
Three platforms tested the last line in fifteen months.
Each date is from the vendor’s own announcement, checked August 2026.
Third-party effects stopped being available across Facebook, Instagram and Messenger, and Spark Studio went with them. Meta put the community at more than 400,000 creators in 190 countries. The effects lived inside Meta’s apps, so Meta’s decision was the whole story.
Discontinued on iOS, Android and Creative Cloud. Content was downloadable until 3 December, after which .real scenes stopped playing for authors and viewers alike. Aero played inside Adobe’s app; removing the app removed the work.
The interesting one, because 8th Wall was WebAR. Logins and export ended in February 2026 and hosted links stop in February 2027. Being on the web did not save it — being on the web is necessary, not sufficient. Someone still has to serve the URL.
What to actually check before committing to any platform: the three questions that predict survival.
A short decision, honestly.
-
1
Is it anchored to something printed?
A poster, a card, packaging, a page. If yes, the browser does this well and the install step is pure loss — you are asking someone holding your artwork to go to a store first.
-
2
Does it need depth, faces or a room?
Placing a model on a floor, anchoring to a face, or persisting across sessions needs native hardware access. That is an app, and no amount of preference changes it.
-
3
Does it have rules?
Scoring, physics, multiplayer — a game engine’s job either way, and the question becomes where it is delivered rather than whether the browser can hold the logic.
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.
If it is a printed thing, try the browser one.
Start from a photo of the piece — the first reveal takes minutes.
Questions, answered.
It is more limited and more reachable. A browser is not given the depth sensor, persistent world anchors or face meshes that ARKit and ARCore expose, so app AR wins on capability. WebAR wins on the step before capability matters: nobody installs anything, which on a printed piece is the difference between being seen and not.
Yes, in the browsers already on the device — iOS Safari and Android Chrome. Image tracking runs on the camera feed on-device. There is no app to install and no store review between finishing the work and someone seeing it.
WebAR, by a wide margin, for a campaign anchored to print. You publish a page and print a code. App AR carries development, two store listings, review cycles and maintenance before a single person sees the experience.
No, and that is the useful lesson. 8th Wall was browser-based and still retired its hosting: logins ended 28 February 2026 and hosted links stop on 28 February 2027. Web delivery removes the store as a single point of failure; it does not remove the host. Ask who serves the URL and what happens to it if they stop.