Another WebXR Milestone: Google Ships an Implementation in Chrome Canary
It's exciting to see us cross another threshold on the way to pervasive Web-based AR/VR: last week at Google IO, Google released an experimental version of WebXR in Chrome.
For me, this has been a long time coming. My research group at Georgia Tech started working on web-based AR browsers when we started the Argon Project back in 2009, and we're now on the fourth version of the browser on iOS and Android (including having a pretty nice open source AR framework called argon.js). WebVR started more than three years ago, with implementations available now in most major browsers. Last year, Trevor Smith and I proposed some initial ideas for a WebXR API that supported both AR and VR, and implemented it in an experimental web framework and browser for iOS (the WebXR Viewer) that we and others are using to do Web-based AR experiements, and explore ideas for future capabilities for WebXR.
If you use VR on desktop or Android, and want to experiment with the evolving WebXR API, check out Chrome Canary and give it a try! They'll be releasing Android AR support in Canary soon, too. I've been waiting until the WebXR Device API is a bit more settled before updating the WebXR API we use in the WebXR Viewer to match it, but I'm looking forward to us and other companies in the Immersive Web Community Group shipping WebXR in other browsers.
WebXR is coming!
The differences between the version of WebXR we proposed last year and the WebXR Device API are small, so updating will be easy. But, I don't want to break experiments people are doing with the WebXR Viewer multiple times, so I've been waiting for the API to be more settled before making the jump. ↩︎
I've been playing around with WebXR/ARcore/ARtoolkit/aframe-ar, AR.js and Argon for a while now, trying to get the right solution for my long-term project. Although I'm getting more and more convinced I can realise it with Web rather than a native app, I'm still deperately missing a solution to recognize logos/images or at least QR codes as markers, like the Vuforia recognition integrated into Argon (just without the custom browser). It's nice to have a Porsche driving around my floor and place astronauts on my desk, but for me a good usage for AR in the web might rather be attaching information to real world products.
This will happen, I'm pretty sure of it. I can't comment on when it will happen in all browsers, but most of the pieces are (or can be) in place. WebXR will start appearing in other browsers beyond Chrome Canary soon, in both the Canary-like-versions and eventually real versions. But probably not till the fall or winter.