BakeAmbientOcclusion - ComfyUI Built-in Node Documentation - ComfyUI

Bakes an ambient-occlusion map from a high-poly mesh into the UV layout of a low-poly mesh. The output is a grayscale image in which white texels are open and dark texels are in crevices; it is meant for the Apply Texture To Mesh node’s occlusion input. Connect the UV-unwrapped low-poly mesh and the high-poly mesh it was decimated from.

Inputs

Parameter	Description	Data Type	Required	Range
`low_poly`	The UV-unwrapped low-poly mesh to bake into. Must have UVs; the node raises an error if they are missing.	MESH	Yes	-
`high_poly`	The high-poly mesh that the low-poly was decimated from, used as the source geometry for occlusion.	MESH	Yes	-
`resolution`	Texture resolution in pixels; each texel receives an occlusion value. Default: 1024.	INT	Yes	64 to 8192 (step 64)
`samples`	Rays per texel. More = smoother, slower. Raise if grainy. Default: 64.	INT	Yes	4 to 1024 (step 4)
`max_distance`	Ray length, as a fraction of the bounding box diagonal. Smaller = tighter, more local occlusion. Default: 0.5.	FLOAT	Yes	0.01 to 2.0 (step 0.01)
`strength`	Scales the occlusion. >1 darkens, <1 lightens. Default: 1.0.	FLOAT	Yes	0.0 to 2.0 (step 0.05)
`bias`	Ray origin lift off the surface, as a fraction of the bounding box diagonal. Raise if even surfaces show dark blotches/holes. Default: 0.01.	FLOAT	Yes	0.0001 to 0.2 (step 0.0005)

Note: low_poly must have UV coordinates — this node never unwraps the mesh. If high_poly contains only one batch item, it is reused for every batch item of low_poly; batch items of low_poly with no faces are skipped and replaced with an all-white image, with a warning logged.

Outputs

Output Name	Description	Data Type
`occlusion`	Grayscale ambient-occlusion image with values in [0,1] (white = open, dark = crevices), one image per batch item of `low_poly`. Intended for the Apply Texture To Mesh node’s occlusion input (packed into the ORM map / occlusionTexture).	IMAGE

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): 63ea6ce5289728d351fdd7d722e9a299ebb1283e1128262a817466ec6d23786a