Location From Point ⚡🅡🅞🅣🅘

This is the manual-entry counterpart to LocationFromMask: instead of deriving a position from a mask, you just type in x, y, and z directly. It's the simplest possible way to get a LOCATION into RyanOnTheInside's proximity/Flex Features system - no segmentation, no mask painting, just numbers.

Why you'd reach for it

Not everything needs to be derived from image data. If you want an effect to react to distance from a fixed point - a spot you've chosen deliberately, not one extracted from a mask - this is the fastest way to define it. It's also the natural starting point when you're testing: rather than building a mask pipeline just to get a LOCATION flowing through your graph, drop this node in, punch in coordinates, and you've got something to wire into LocationTransform or any other proximity node while you build the rest of the workflow around it.

batch_count is worth calling out specifically: it lets you generate multiple copies of the same point in one go, which matters when the rest of your graph is working on a batch of frames and expects a location per frame rather than a single static one.

The inputs and output

• x / y (required, FLOAT, default 0, minimum 0) - the horizontal and vertical position.
• z (required, FLOAT, default 0, clamped 0.0–1.0) - depth. Note the range is different from x/y: it's normalized 0–1 rather than open-ended, so treat it as "how far into the depth range" rather than a pixel coordinate.
• batch_count (required, INT, default 1, minimum 1) - how many copies of this same point to output, for matching a batch of frames.

Output is a single LOCATION, ready to feed into LocationTransform (to animate it with a Flex Feature) or wherever else the pack expects a location.

How to install it

The README came back empty when I checked for install specifics, so the standard route applies here too:

ComfyUI Manager: search "RyanOnTheInside", install, restart.

Manual:

cd ComfyUI/custom_nodes
git clone https://github.com/ryanontheinside/ComfyUI_RyanOnTheInside

Restart ComfyUI. It's a large pack spanning audio, mask, and proximity systems, so expect a slightly longer first install than a narrow single-purpose node pack.

Troubleshooting

LOCATION only connects to other RyanOnTheInside nodes - it's a pack-specific type, not something you can wire into a generic ComfyUI coordinate input. If a downstream node behaves as if z is always at one extreme, remember it's clamped 0–1, unlike x and y which are open-ended - a z value you meant as "a pixel coordinate of 500" will just clamp to 1.0 and do nothing like what you expected. If you're feeding a batched workflow and something errors on a batch-size mismatch, check batch_count first - it needs to match (or be broadcastable to) whatever frame count the rest of your graph is working with.