This node can perform several different functions according to the Operation parameter.

The common operations are:

• Boolean operations (union, intersect, subtract) between two "solid" models:

  

• Shattering a solid model using cutting surfaces:

  

  

  You can shatter with any number of complex cutting surfaces, allowing natural-looking, artistically controlled destruction.

• Generating polylines along the seams where two models intersect.

• Detecting intersections and putting intersecting polygons into groups.

You can specify the input models as "solid" (treat the model surface as if it encloses a solid filled space) or "surface" (treat the model surfaces like flat planes).

In the images below, the "shatter" examples use an Exploded View node to make the separate pieces clear.

Solid/Solid boolean operations

Before	To illustrate boolean operations between solids, we’ll use two simple offset boxes.
Union	Creates a new solid which encloses the combined volume of the two inputs.
Intersect	Creates a new solid which encloses any shared volume(s) between the two inputs.
Subtract	Removes the shared volume(s) from one or both inputs.
Shatter	Like a combination of Intersect and Subtract: cuts along the intersections between the surfaces of the solids to carve out new shapes.
Seam	Outputs polyline(s) where the surfaces of the two solids intersect.

Solid/Surface boolean operations

Before	To illustrate boolean operations between a solid and a surface, we’ll use a simple box and a corner.
Union	Combines the solid with double walls around parts of the surface outside the solid’s volume. Note that the open surfaces become double-walled, or what you might think of as "solids with zero volume" and overlapping points. In the image on the right, the parts of the "corner" outside the cube are now double-walled with double points at each corner (see the schematic views below). This is a logically consistent outcome given how the node sees the interaction between solids and surfaces, but it can be surprising if you don’t expect it. You might be able to use this geometry for what you want, but in general this operation is not very useful.
Intersect	Trims away any parts of the surface that are exterior to the solid.
Subtract	Subtracting a surface from a solid creates double-walled cuts in the solid where it intersects the surface. If the surface goes "all the way through" the solid, this is like shattering, where the surface cuts the solid into disconnected pieces. However, if (as on the right) the surface doesn’t fully intersect the solid, it "cuts out" a zero-width shape from the solid, leaving double walls with double points at the corners (see the schematic views below). Subtracting a solid from a surface trims any parts of the surface that are inside the solid.
Shatter	Like a combination of Intersect and Subtract: carves out the intersections into new shapes. If the surface goes "all the way through" the solid, the surface cuts the solid into disconnected pieces. However, if the surface doesn’t fully intersect the solid, it "cuts out" a zero-width shape from the solid, leaving double walls with double points at the corners (see the schematic views below).
Seam	Outputs polyline(s) where the surface intersects the solid.

Surface/Surface boolean operations

Before	To illustrate boolean operations between surfaces, we’ll use two corner shape. The two shapes have an overlapping surface (in the top right) as well as an intersection (bottom).
Union	Combines coincident surfaces, and combines intersecting parts of the two surfaces.
Intersect	Only retains coincident surfaces.
Subtract	Removes coincident surfaces, and creates double-sided cuts along intersections.
Shatter	Like a combination of Intersect and Subtract: carves out intersections, and retains coincident surfaces as new shapes.
Seam	Outputs polygons for coincident surfaces, and polylines where the surfaces intersect.

For solid models with "overlapping" or "concentric" surfaces (for example, boxes within boxes, or geometry with self-intersections), you can use the "Custom" operation to extract only the solid area at a certain "depth" (with respect to either or both inputs).

The "depth" of any point inside the solid is measured as the minimum number of surfaces you'd have to cross to get outside the model.

In this illustration, a model with three concentric boxes (gray) is combined with a single-level box (green). The numbers in the illustration show the "depth" of each volume. So, if you set the custom range to 4-9999, the output would only contain the dark-gray square in the center.

In this illustration, a single model has a two self-intersections, creating a nested space inside. Setting the custom range to 2-9999 would delete the light grey parts and leave the dark grey box. The Custom operation can be useful for removing the "outer layers" to get at these kinds of spaces inside.

This mode passes through the first input geometry, but adds optional groups and/or attributes containing the intersecting polygons. This lets you create effects using the intersecting polygons.

This mode is used for internal testing and is probably not generally useful, except possibly as a quick tetrahedralize. It outputs the raw triangulated mesh from an intermediate step in the node’s algorithm, equivalent to the union of all polygons as surfaces. The polygons may have inconsistent orientations.

• In the parameters you specify whether an input is a "solid" or "surface" (zero-width). If you specify that an input is "solid", you need to make sure it’s a good solid.

  • In general solids should be closed and airtight (but see below).

  • Solids should have all normals pointing outward (consistent polygon winding directions). You can turn on display of normals in the viewport to check for inverted or double-layered polygons.

    

  • Solids should not have non-manifold edges (that is, they should have no edges shared by more than two polygons, like a T junction).

  If the node encounters non-manifold edges or inconsistent normals, it will do something with the geometry, but it might not be what you wanted (the problems can give the algorithm the wrong idea of what’s "inside" and "outside"). In these cases the node will show a warning explaining what created the ambiguity. You can read the warning in the node info window. The warning includes a group spec specifying the problem polygons.

• It is possible to use open surfaces as solids as long as this does not create logical contradictions. For example, no unshared edge of a mesh treated as solid can cross the interior of that or any other solid.

  If you try to use an open surface as a solid and it creates inconsistencies, the node will highlight the problem boundary edges in red.

• You can turn off visualization guides when the node is active. Right-click in the viewport or click the Boolean icon on the left side of the operation toolbar (at the top of the viewer), and turn off visualization options in the menu.

• There are different ways to visualize connectivity after the boolean operation:

  • Use the group list with the Select tool in the viewport to highlight disconnected parts as you move the mouse over them. This doesn’t require modifying the network.

  • Use an Assemble node to give the polygons of each disconnected piece a common name attribute value, then use a Color node or color visualizer with the "Random color by attribute" option to color the polygons based on the name value:

    

  • Use an Exploded View node to move the parts away from each other.

• You can connect two polygonal geometries to the node’s two inputs, or connect one input and operate on different groups using the Set A Group and Set B Group parameters.

  If you connect only one input and leave Set A and Set B groups blank, the boolean operation applies to the model with itself. You can use this with Shatter to separate the model along its self-intersections. The other boolean operations are logically consistent (union and intersect pass everything through, subtract deletes everything) but not very useful.

• This node interpolates point/vertex attributes and copies primitive attributes.

• It’s usually not worth trying to align edges on the two models for aesthetic purposes. For example, trying to put a cutting plane exactly on the equator of a sphere. Due to tiny numeric precision errors, the two edges can look lined up even zoomed in but technically have a gap between them according to the computer. Boolean will automatically clean up tiny polygons created by this kind of mis-alignment.

• Red dots in the node’s visualization show problem areas, such as microscopically tiny polygons in the output.

• If you're doing highly complicated procedural booleans, for example with extremely fine details/differences between surfaces (as opposed to using booleans for visual modeling), it’s possible to get unintended results if you follow a boolean with another boolean, because precision issues¹ can cause microscopic self-intersections in Boolean’s output.

  You can try turning off Assume seam polygons are flat on the upstream Boolean node. If that doesn’t work, set Detriangulate to "Only unchanged polygons". These settings leave more triangles in the output but might fix the self-intersection.