VECTOR3 (JOV)

VECTOR3 (JOV) bundles three numbers into a single VEC3 wire, and that's the entire job. If you're coming from a world where you just type "1.0, 2.0, 3.0" into a widget, this looks pointless - but in a graph you can't always type, because the values are often animated. When a WAVE GEN is driving your X and you want Y and Z to track it, you need a node that takes three inputs and puts them on one output, and that's this node.

When you'd actually use it

The honest answer: less often than its 2D and 4D siblings. VEC3s show up in 3D-flavored work - RGB colors, 3D coordinates, normals, the width/height/channel trio some nodes want - and Jovimetrix's own ecosystem leans VEC2 and VEC4. You'll reach for this when a node asks for an RGB-ish triple and you want to drive the channels separately, or when you're doing anything spatial and a single wire keeps your graph readable. It's also ComfyMath-vector compatible, so the VEC3 you build can flow into ComfyMath's vector math nodes.

Inputs and outputs

Three numeric inputs - X, Y, Z (each labeled "1st/2nd/3rd channel value") - accepting booleans, floats or ints over the full signed 64-bit range. Leave any unconnected and it contributes 0. Outputs mirror the 2D node:

• VEC3 - the triple as floats.
• VEC3INT - the triple rounded to integers, for when downstream wants whole numbers.

Install

It comes with Jovimetrix: git clone https://github.com/Amorano/Jovimetrix.git into ComfyUI/custom_nodes, then pip install -r requirements.txt (Windows portable: .\python_embed\python.exe -s -m pip install -r requirements.txt), or search "Jovimetrix" in ComfyUI Manager. No model downloads.

The only real gotcha is remembering what a VEC3 is in your pipeline. Jovimetrix treats vectors as plain tuples of floats, so a VEC3 and a three-component VEC4 are different types - if a node rejects your VEC3, check whether it wants a VEC4 with a fourth (often 1.0) component rather than assuming a mismatch. Otherwise: build a triple, wire it, done.