Open-Source WebAR vs Stories AR: Build or Buy

What open source actually gives you

MindAR and AR.js are JavaScript libraries for browser-based image tracking. They are genuinely free (open licenses), actively used, and capable of exactly the core trick: recognize a target image through the camera and anchor content on top of it. If you are a developer, you can have a demo running on your laptop in an evening.

What they give you is the tracking engine — nothing more. There is no viewer product, no content system, no analytics. That is by design: they are libraries, building blocks for developers.

For some projects that is precisely the appeal: you can build interactions no platform offers, integrate AR into an existing app or site, and never pay a subscription.

What you build yourself around a library

Between 'demo on a laptop' and 'product a client scans at a wedding' sits a list of engineering work. A viewer UI: camera permission flow, loading states, fallbacks when tracking is lost. Video playback anchored to the target — with the mobile-browser autoplay and codec quirks that come with it, across Safari on iOS and Chrome on Android.

Then marker preparation: image targets must be compiled per photo, and someone must judge which photos track well. Then hosting: the page, the videos, a CDN if clients are not all in one city. Then the operational layer — when a client wants to swap a video, that is a redeploy unless you have built a dashboard, which is its own project.

None of this is exotic work. It is just work — ongoing, browser-update-sensitive, and entirely yours. The recent 8th Wall shutdown timeline (access closed February 2026, hosted projects stopping February 28, 2027) is a reminder that even paid infrastructure changes; with self-hosting you carry that responsibility yourself, for better and worse. If you are comparing platforms too, see the 8th Wall alternative breakdown.

What the no-code path gives you

Stories AR packages that whole stack as a product. You sign up at pro.stories-ar.com, upload a photo + video pair, and get a permanent link and QR code. The viewer runs in the browser — Safari, Chrome — with camera processing done locally on the device and nothing recorded or uploaded. How the underlying WebAR approach works is covered on the landing.

The operational layer is the real difference. A dashboard lists all your AR photos; videos are replaceable anytime without changing the QR code or reprinting; batch upload handles an album or a calendar in one session. More than 62 000 photographers use the platform and have created 259 000+ AR photos — the workflows around print products are the part that took years, not the tracking.

Pricing: first AR photos free, PRO at $24/month billed yearly with 20 AR photos per month, and white-label plans for agencies and photo labs that want the viewer under their own brand.

When open source is the right call

Choose the build path if any of these is true: you have in-house developers and the AR experience is part of a bigger product; you need custom interactions beyond photo-plays-video (games, 3D scenes, world tracking — where you would look at Zappar, Blippar or native ARKit/ARCore anyway); you have hard data-residency or self-hosting requirements; or the economics of a very large, long-lived deployment favor owning the stack.

Also legitimate: you enjoy it. Plenty of good tooling exists because a developer wanted control. If AR is your product rather than an add-on to your photography, owning the code is a defensible strategy.

When no-code wins

Choose the buy path if you sell photos, prints, albums or packaging and AR is a feature of your product, not your product. The relevant metric is time from client order to working QR code — minutes on a platform versus a development project on a library.

The hidden factor is iteration: clients change their minds, videos get replaced, seasonal products get refreshed. On a platform that is a dashboard click. On a self-built stack every change touches a developer.

A fair test costs nothing: build one photo on the free tier, and if you are technical, spend an evening with MindAR. Compare not the demos but the path to your tenth client order.