Hyper3d Rodin V2 Image To 3d
Playground
Try it on WavespeedAI!Hyper3d Rodin v2 is a powerful image-to-3D model that creates high-quality 3D assets from single image.
Features
Hyper3D-Rodin-Gen-2
Hyper3D-Rodin-Gen-2 is a commercial 3D generation system from Hyper3D that converts text prompts and 2D images into production-ready 3D assets with UVs and textures, targeting games, film, XR, product design, and 3D printing workflows.
Technical Highlights
The system provides a unified pipeline for 3D geometry + texture generation:
-
Gen-2 Large-Scale 3D Diffusion Model:
Utilizes a new large 3D diffusion backbone (BANG-style architecture) trained on large-scale 3D data to produce cleaner topology and higher-fidelity meshes compared with earlier generations. -
Unified Text-to-3D & Image-to-3D API:
A single/api/v2/rodinendpoint supports both text prompts and images:- Text-to-3D when only
promptis provided - Image-to-3D when one or multiple images are uploaded
Multi-image modes (e.g., feature fusion, multi-view) help reconstruct more accurate shapes from different viewpoints.
- Text-to-3D when only
-
Mesh & Texture Output for Production Pipelines:
Directly outputs UV-unwrapped meshes plus textures, supporting formats such asglb,usdz,fbx,obj, andstl. Texture modes include PBR maps (albedo, normal, roughness, metallic, etc.) and shaded/baked textures for stylized looks. -
Topology and Polycount Control:
Provides quality presets (e.g., extra-low, low, medium, high) that map to different face counts for triangle (Raw) or quad (Quad) meshes. Advanced users can override presets with explicit face-count settings, and optional add-ons can further increase polygon density and texture resolution. -
Geometric Control and Posing:
Supports ControlNet-style geometric constraints via 3D bounding boxes to control the approximate width, height, and length of the generated object. Character-generation options (such as enforcing T/A-pose) simplify downstream rigging and animation.
Performance and Efficiency Optimization
-
Async Cloud Generation with Tiered Latency:
Uses an asynchronous cloud pipeline: submit a job, poll status, then download results. Different quality tiers correspond to different latencies, allowing a trade-off between speed and detail. -
Cost & Quality Trade-Off via Tiers and Add-Ons:
A credit-based billing scheme lets users balance cost, quality, and speed. Lower tiers and fewer add-ons yield faster, cheaper generations suitable for rapid iteration, while higher tiers and enhancement packs target final-quality assets. -
Multi-Modal Input and Format Interoperability:
Supports pure text, single image, and multi-view image inputs within the same API. Output formats are compatible with common DCC and game engines such as Unity, Unreal, and standard 3D-printing workflows via STL. -
Developer Ecosystem and Integrations:
Offers official documentation, minimal code samples, and additional endpoints (for checking balance, status polling, and texture-only operations). Rodin-Gen-2 is also available through several 3rd-party platforms, making integration into existing production pipelines and low-code tools straightforward.
Parameters
- material
- Type:
string· enum - Default:
PBR - Possible values:
PBR,Shaded,All
- Type:
Authentication
For authentication details, please refer to the Authentication Guide.
API Endpoints
Submit Task & Query Result
# Submit the task
curl --location --request POST "https://api.wavespeed.ai/api/v3/hyper3d/rodin-v2/image-to-3d" \
--header "Content-Type: application/json" \
--header "Authorization: Bearer ${WAVESPEED_API_KEY}" \
--data-raw '{}'
# Get the result
curl --location --request GET "https://api.wavespeed.ai/api/v3/predictions/${requestId}/result" \
--header "Authorization: Bearer ${WAVESPEED_API_KEY}"
Parameters
Task Submission Parameters
Request Parameters
| Parameter | Type | Required | Default | Range | Description |
|---|---|---|---|---|---|
| prompt | string | No | - | A textual prompt to guide the model generation. | |
| images | array | Yes | [] | 1 ~ 5 items | Images to be used in generation, up to 5 images. As the form data request will preserve the order of the images, the first image will be the image for material generation. |
| material | string | No | - | PBR, All, Shaded | The material type. |
| quality_and_mesh | string | No | - | 4k_Quad, 8k_Quad, 18k_Quad, 50k_Quad, 2K_Triangle, 20K_Triangle, 250K_Triangle, 500K_Triangle | The generation quality and mesh mode. |
| geometry_file_format | string | No | - | glb, fbx, obj, stl, usdz | The format of the output geometry file. |
| addons | string | No | - | HighPack | Generate 4K resolution texture instead of the default 2K. If Quad mode, the number of faces will be ~16 times of the number of faces selected in the quality parameter. |
| bbox_condition | array | No | - | - | This is a controlnet that controls the maxmimum sized of the generated model. |
| TAPose | boolean | No | - | - | Control the generation result to T/A Pose. |
| use_original_alpha | boolean | No | - | - | Used when processing the image. |
| preview_render | boolean | No | - | - | Provided in the download list. |
| seed | integer | No | - | -1 ~ 2147483647 | Seed for random number generator. Set to 0 to use a random seed. |
Response Parameters
| Parameter | Type | Description |
|---|---|---|
| code | integer | HTTP status code (e.g., 200 for success) |
| message | string | Status message (e.g., “success”) |
| data.id | string | Unique identifier for the prediction, Task Id |
| data.model | string | Model ID used for the prediction |
| data.outputs | array | Array of URLs to the generated content (empty when status is not completed) |
| data.urls | object | Object containing related API endpoints |
| data.urls.get | string | URL to retrieve the prediction result |
| data.has_nsfw_contents | array | Array of boolean values indicating NSFW detection for each output |
| data.status | string | Status of the task: created, processing, completed, or failed |
| data.created_at | string | ISO timestamp of when the request was created (e.g., “2023-04-01T12:34:56.789Z”) |
| data.error | string | Error message (empty if no error occurred) |
| data.timings | object | Object containing timing details |
| data.timings.inference | integer | Inference time in milliseconds |
Result Request Parameters
| Parameter | Type | Required | Default | Description |
|---|---|---|---|---|
| id | string | Yes | - | Task ID |
Result Response Parameters
| Parameter | Type | Description |
|---|---|---|
| code | integer | HTTP status code (e.g., 200 for success) |
| message | string | Status message (e.g., “success”) |
| data | object | The prediction data object containing all details |
| data.id | string | Unique identifier for the prediction, the ID of the prediction to get |
| data.model | string | Model ID used for the prediction |
| data.outputs | string | Array of URLs to the generated content (empty when status is not completed). |
| data.urls | object | Object containing related API endpoints |
| data.urls.get | string | URL to retrieve the prediction result |
| data.has_nsfw_contents | array | Array of boolean values indicating NSFW detection for each output |
| data.status | string | Status of the task: created, processing, completed, or failed |
| data.created_at | string | ISO timestamp of when the request was created (e.g., “2023-04-01T12:34:56.789Z”) |
| data.error | string | Error message (empty if no error occurred) |
| data.timings | object | Object containing timing details |
| data.timings.inference | integer | Inference time in milliseconds |