RenderMesh - ComfyUI Built-in Node Documentation

This node renders a 3D mesh into a 2D image by ray-casting a single view.

This node renders a 3D mesh into a 2D image by ray-casting a single view. It can output the textured mesh, vertex colors, a solid shaded surface, surface normals, or depth. The camera and optional model transform can come from a Load3D / Preview3D viewer; if no camera is connected, a default front view is auto-framed.

Inputs

Parameter	Description	Data Type	Required	Range
mesh	The 3D mesh to render.	MESH	Yes	—
mode	What to render. auto: texture if present, else vertex colours, else shaded clay. (default: “auto”)	COMBO	Yes	"auto" "texture" "vertex colors" "solid" "normal" "depth"
width	Width of the rendered image in pixels. (default: 1024)	INT	Yes	64 to 4096 (step 8)
height	Height of the rendered image in pixels. (default: 1024)	INT	Yes	64 to 4096 (step 8)
background	Background color used for pixels the mesh does not cover. (default: “#000000”)	COLOR	Yes	—
model_3d_info	Model transform from the same Load3D / Preview3D viewer. Connect it with camera_info to match the viewer framing.	LOAD3D_MODEL_INFO	No	—
camera_info	Camera from a Load3D / Preview3D viewer or a Create Camera Info node. If none is connected, a default front view is auto-framed.	LOAD3D_CAMERA	No	—

Note: Only the first item of a batched mesh is rendered — if the mesh batch contains more than one item, the node logs a warning and uses the first one. The texture mode requires the mesh to have both a texture and UVs, and the vertex colors mode requires vertex colors; if the data for the selected mode is not available, the node falls back to the solid shaded render. The normal mode uses the mesh’s smooth vertex normals when present, otherwise the per-face normal. In depth mode, closer surfaces appear brighter (near = white) and pixels not covered by the mesh stay black. The model_3d_info and camera_info are meant to be connected together from the same Load3D / Preview3D viewer so the render matches the viewer framing.

Outputs

Output Name	Description	Data Type
image	The rendered image of the mesh.	IMAGE
mask	A mask that is 1.0 where the mesh was rendered and 0.0 elsewhere.	MASK

This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! Edit on GitHub

---

Source fingerprint (SHA-256): d23e85a904520eb2dfed899eb3e6a9cf45c980df00c034503687ac4eccc66ac4