Meshy 6: Image to Model
Upload an image and generate a 3D model with Meshy 6. Create characters, objects, or mechanical parts with production-quality geometry and clean topology.
简介
This ComfyUI tutorial workflow shows you how to turn a single 2D image into a production-ready 3D asset using the Meshy 6 API. Starting with LoadImage, your reference is sent to MeshyImageToModelNode, which generates a clean, watertight mesh with UVs directly from the image. You can then iterate on quality and detail using MeshyRefineNode, add PBR textures with MeshyTextureNode, and view results in real time through the Preview3D node.
你需要提供
- 正在读取工作流
在 ComfyUI 中运行。 此工作流无法在本页面运行。你可以在 Comfy Cloud 中打开（已预先加载所需内容），或下载后在自己的 ComfyUI 中使用。
在 Cloud 中试用常见问答
Yes. The MeshyImageToModelNode (and related Meshy nodes) call the Meshy 6 API. Provide a valid API key in the node(s) or your environment so jobs can run and return results.
By default, models are auto-saved to ComfyUI/output/ via the SaveGLB node. You can also export directly from the Preview3D node’s export button to download the GLB.
Start with a high-quality image showing a clear silhouette of a single subject. After the initial MeshyImageToModelNode pass, run MeshyRefineNode to improve geometry, then MeshyTextureNode for PBR textures. Iterate while checking results in Preview3D.
MeshyRigModelNode works best with humanoids and simple bipeds. Props and mechanical parts typically don’t need rigging; skip rigging for those. Use MeshyAnimateModelNode to preview motion where appropriate.