Rodin now rigs creatures and edits the skeleton on request

Automatic rigging has been added to the 3D model generator Hyper3D Rodin. It works with humans, animals and fantasy creatures and places bones down to individual fingers. The skeleton can be edited with text: a command like “Add tail bones” or “Add ear bones” adds tail or ear bones. The update was released on November 3, 2026. The rigging entry appeared in Rodin's changelog back on October 23, and now prompt-based skeleton editing and motion presets have been added to it.

For anyone who builds characters through generation, this shortens the path to animation. Before rigging arrived, a static model from Rodin had to be exported and rigged in another tool. Now the generated creature can be set in motion without leaving the service.

“Need a different skeleton? Just describe it. Type Add tail bones or Add ear bones to adjust the rig with a prompt.”

That is how the new capability is described in Rodin's changelog.

From skeleton to first animation in a single window

The workflow looks like this. The user gets a model, and the service automatically builds a rig for it. If the character needs a tail, ears or other non-standard parts, the missing bones are added with a text command. Then you can pick a motion preset and tweak the animation with interactive controls.

• Character types: humans, animals, fantasy creatures.
• Detail: the skeleton goes down to individual fingers.
• Rig editing: via text prompt.
• Animation: motion presets with subsequent fine-tuning.

Some questions are left open in the changelog. It does not specify what pose the model should be in before rigging or what mesh topology is suitable for it. It is also unclear where the limits of a prompt-defined skeleton lie. There is no information on whether the Rodin DCC Bridge plugin passes the skeleton and animation to third-party packages or only geometry with textures.

What else has changed in Rodin since May

Rigging caps off six months of dense updates. The biggest of them is the Rodin Gen-2.5 model, released on May 24. It can split a model into parts manually via the BANG to Parts tool, refine details locally with the Part Refine function and output up to 10 variants in a single batch generation. For a modeler, judging by the description, that means being able to rework individual parts of an object without regenerating the whole thing. Since August 29, AI segmentation has complemented manual splitting, and Voxel ControlNet settings have given more precise control over geometry.

The July 2 version of Gen-2.5 improved results in common categories: characters, weapons and vehicles. The Thinking Effort High mode was significantly improved. On September 13, Gen-2.5 Agent-Ready was released with a second generation of the native 3D texture architecture, improved support for 8K and 12K textures and full support for agentic workflows.

Many changes concerned textures:

• 12K textures have been available in beta since June 7, and 8K was opened to all users on July 2;
• HD and 8K toggles appeared in the Texture panel;
• the Partial Redo function arrived in Texture Stage (since July 31);
• Smart UV in Texture Stage (since October 12);
• Emissive and Bloom preview, a separate Texture Viewer and magnified detail inspection.

Several changes matter for export and integration. Rodin DCC Bridge gained Gen-2.5 support, the Unreal Engine add-on started working on macOS, and on September 24 an integration with ZBrush appeared. Agentic Mode and the Hyper3D CLI command line shipped alongside it. Export to 3MF for models with color quantization was added to the Pack panel — probably with color 3D printing in mind. On August 29, the WorldGen scene generator with camera control and export was opened to all users.

Who benefits and where the catch is

The “generator plus a separate rigger” combination existed before. In a Creative Bloq guide, for example, a model from Rodin was taken to animation through AccuRIG and ActorCore, with subsequent work in iClone. Built-in rigging probably removes at least one tool from that chain. The biggest winners may be indie developers and the authors of clips and prototypes who need a moving character fast and without production-level requirements.

There is also a difference from Meshy Rigging. Judging by the Fuser documentation, that rigger is designed for humanoids with clearly defined limbs and also offers a library of animation presets. Rodin claims support for animals and fantasy creatures, and non-standard bones are added with a prompt. For creatures with tails or extra ears that is an advantage on paper, if the quality of the rig holds up under scrutiny.

For riggers in studios, the new feature is unlikely to replace manual work. The changelog says nothing about skinning weights, controllers or compatibility with studio standards, and the question of passing the skeleton through DCC Bridge remains open. The price is unknown too: rigging is not mentioned separately on the pricing page. The free plan offers pay-per-result with credits at $1.5 and 10 private assets. The Creator plan costs $30 a month or $24 with annual payment ($288 a year) and is designed for around 60 models; it lists redo limits — 20 for geometry and 6 for material. Business at $120 a month or $96 with annual payment ($1,152) gives full API access and about 416 models. How many credits a rig and its recalculation consume is not yet clear.

What is more useful to you in Rodin's rigging?

• Skeleton down to the fingers
• Tail bones