Presentation stills for Houdini Core assets and Solaris USD views.
Presentation stills from Houdini Core procedural assets, character poses, game-art variations, and Solaris USD lookdev views.
Presentation stills from Houdini Core procedural assets, character poses, game-art variations, and Solaris USD lookdev views.
Houdini Core suits modelers, lighters, riggers, animators, and game artists who need to iterate without discarding the node networks behind an asset. A parameterized prop, modular environment piece, character pose, motion-graphics form, or Houdini Digital Asset can produce many credible review states from one procedural setup. For layout, lighting, and lookdev, the same work can move into Solaris, where LOP networks build and edit a USD stage with cameras, materials, lights, variants, and assembled assets.
RENDERCAD is useful when the team needs a fast presentation still from the clearest working view: a well-framed Scene View capture for a modeled or rigged asset, or a Solaris viewport capture that makes the chosen camera, composition, material direction, and scale readable. Use it to communicate intended mood or context, compare visual directions, and make an internal review, client update, pitch, or asset presentation easier to discuss. It is not a replacement for the Houdini scene, USD layers, or a final Karma render. Keep procedural controls, USD composition, exact materials, lighting, and production-render decisions in Houdini Core as the source of truth.
First, set the procedural parameters, asset variant, or animation frame you want to discuss. In Solaris, verify the active USD stage and select a camera or viewport angle that clearly shows the layout, material assignments, lighting direction, and useful scale references. Hide node panes, guides, handles, selection outlines, and temporary helpers before capturing the cleanest view.
Upload that capture to RENDERCAD to create a presentation still for an asset approval, lookdev conversation, game-art review, or marketing concept. Compare the result against the Houdini view, then return to Core to adjust the controlling nodes, camera, USD layers, materials, or final Karma settings. Use the generated image to guide discussion, not to validate geometry, animation, compositing, or renderer-specific output.
Modeled by CADSharp in Onshape. Rendered with RENDERCAD.
Add | Change | or Remove referenced objects with the stroke of a brush.
RENDERCAD uses a maintained selection of image-generation models for different rendering tasks. Results can vary by model and input, so compare generated output with the source design before relying on geometry, materials, or dimensions.
| Feature | RENDERCAD | Houdini Core rendering workflow |
|---|---|---|
| Engine | Generative AI | Houdini GL and, in Solaris, Karma CPU or XPU and available Hydra render delegates |
| Output | Images + Video | Still images, animation frame sequences, render gallery snapshots, and render passes |
| Resolution | HD / 4K | Configured through Houdini render settings and limited in practice by the scene, selected renderer, and available hardware |
| Image Speed | Image renders complete in 60 seconds. | Scene-dependent; complex USD composition, instancing, volumes, materials, and sampling can extend render time |
| Video Speed | Videos complete in 2-5 minutes. | Animation is rendered over a frame range; speed depends on the scene, cache state, renderer, sampling, and hardware |
| Scene Set Up | No separate local scene | Procedural geometry and asset networks, cameras, lights, materials, render settings, and Solaris USD stages built with LOP networks |
| Input | Screen Snip | Houdini .hip scenes, SOP geometry, HDAs, USD stages, cameras, lights, materials, and render nodes |
| Hardware | Cloud GPU | Houdini desktop workflow, with optional render-farm or pipeline tooling |
| Install | None | Houdini Core installation; additional render delegates and render-farm tools may need separate configuration |
| Cost | from $9/mo | Requires suitable Houdini licensing; some optional render tools may have separate licensing |
Read the numbers. We trust you to know theirs.
Skip a separate local scene build. Pick a condition, material, and background while RENDERCAD handles the image-generation step. Image renders complete in 60 seconds.
™
RENDERCAD™
Use the browser workflow directly, capture a selected view with the desktop app, or start from the browser extension.
Get StartedCommon solutions for Houdini Core users
The core workflow is available in a supported browser and accepts a selected 2D image, so it does not require a rendering plugin in the source application.
The model can reinterpret curves and surface transitions from the 2D input. Compare the result with the source because edges and geometry can drift.
It accepts a clear viewport screenshot, so a separate local material, light, and camera scene is not required for the image-generation workflow.
It lets you describe motion, steam, lighting changes, or atmosphere for a generated clip. Added details may not exist in the source model.
Yes, if you are authorized to use the photo. It can guide the environment, while placement, lighting, and shadows remain generated interpretations.
Yes. Draw a mask around the product and direct the background region separately. Inspect the full result because unmasked content can still drift during generation.
Yes. Mask a region, then assign a specific reference image just to that masked area. Other areas can have different references or default to the prompt.
Capture a clear view, describe the leather type and finish, and use a material reference when surface direction matters. Review the generated depth and stitching rather than treating them as product specifications.
Choose a suitable background and describe a reflective floor. Inspect the generated reflection for alignment, product fidelity, and unintended details.
Choose the Orbit camera preset and keep the source image as the authority for product shape and visible details. Generated parallax and newly revealed areas can vary.
If the source viewport is visible but a local renderer fails, capture a clear screenshot and use it as a RENDERCAD input. Image renders complete in 60 seconds.
Yes. Upload a photo of a foam model and prompt "Finished plastic product" to visualize the final look.