Hops in iRhino3D WIP

iRhino3D WIP (8.0.23123 (11.1.1) - May 3) available via Testflight can solve hops definitions on a remote server now.

This is an unfinished feature with some known issues and many unknown ones. It may also crash on you unexpectedly, so use at your own risk :slight_smile:

For serious work, always use the Appstore version of iRhino3D.

Configuring the Server

You need a Compute server over HTTPS in order to use this feature. You can find step by step instructions to set up a compute server here.

  • Once you have a compute server running, enter the server (Https) url and API key by going to Settings > Extra > Hops

  • Tap “Check Connection” to ensure the server is running.
  • You can now upload a Hops definition and display the results

Known Limitations

  • Supported inputs: numbers, text, booleans
  • Supported outputs: breps, curves, points, meshes, subd

This is great @mkarimi! Love the idea of being able to use iRhino3D as a client facing configurator app

1 Like

@mkarimi - can you elaborate on the HOPS-iRhino3d integration roadmap?

@mkarimi this looks so promising, may I ask do you have the gh file for the bench so I can test, I’d love to show the powers that be at our company these different ways of presenting designs to clients, tia

also in the latest build my models are coming in way off origin, super hard to find ! The size of the grid seems really small, or tt may be a scaling issue ? Is there a way to zoom to selected object or layer ?. I imported the model in from OneDrive on my iphone.

We don’t have a roadmap, there are some obvious things that we need support, for example all possible input and output types. But besides that we want to hear from users what they think the roadmap should be like

Here is the bench model that I used in that video. Note, you’ll need to configure your remote VM instance running Rhino.Compute to accept HTTPS. To do that, you can follow this guide. Rhino - Configure Compute to use HTTPS

Bench.gh (33.0 KB)


Are you talking about model generated with Hops in iRhino or simply opening a file in iRhino?
In the latter case, iRhino doesn’t change the scale/origin of your model. Hops integration in iRhino is units agnostic, it’s possible you need to adjust the document scale before running hops.

There’s no zoom selected ATM, but double tap anywhere on the screen to “zoom to extents”

yes I was hoping for zoom selected, or highlight a layer and zoom to the object on the layer, or something similar

1 Like

@AndyPayne ooooo sounds a lot of faffing :crazy_face:, I’d have to check with our IT consultants about this, they get upset when you do things to the ‘system’ :laughing:

That makes sense. I can get this in the next WIP build.


I’d like some more doco and working examples.

I’ve been following a couple of youtubers on how to get things to work, but it seems pretty sparse out there.

How do we do the setup? What is the scope of the iRhino3d capability with HOPS? Do we need Compute on Windows in EC2/Cloud? Why not serverless? etc.

For using Hops in iRhino3d you need compute server running in the cloud. We have step by step instructions for setting up a Compute server on Azure and AWS.

Once you have that working, you can enter the serve URL and API key in iRhino and solve a Hops definition remotely.

thanks @mkarimi !

yeah the concept doesn’t seem right. If we need a Windows EC2 instance running 24/7 it would appear that Rhino compute isn’t really designed for Cloud.

Setting up a Rhino compute server on Azure/AWS/GCP requires an OS level & app level patching regime whence the pets/cattle debate arises and the question becomes the “why isn’t Rhino compute a 12 factor app?”

Is it possible for HOPPS to go serverless, or at least show how to containerise Rhino to run in ECS for example?

The HOPPS-GH-RHINO Cloud Architecture needs some thinking through with doco and examples.

We are at a very early stage of prototyping ideas on different ways to use iRhino. What we currently have for offerings with regards to compute is what Morteza provided in the links above.

There is no current serverless option for Rhino.Compute. You would need to launch and manage a Windows EC2 instance.

1 Like

Thanks @stevebaer

I wonder whether it would be better to focus on WebGPU than an EC2-based compute concept?

In design terms, EC2’s are a backwards step as mentioned above.


1 Like

if I have a NAS can it work?

I assume you mean running the compute server on your NAS machine?
As long as you can connect to it via Https it should work.