Index Wrap / Clamp

Index Wrap / Clamp is a two-behavior utility that solves one recurring annoyance in ComfyUI: keeping an integer inside a valid range. You hand it an index and a size, and it either wraps the index around (like a modulo) or clamps it to the edges. That's it. It's the kind of node you don't think about until you're three layers deep in an animation loop and the tenth frame sends index 12 into a list that only has five entries.

How it works

The math is deliberately minimal:

• If size <= 0, it returns 0 - it refuses to divide by nothing.
• With wrap on, output is index % size. Python's modulo is always non-negative, so you get a value in [0, size-1], no manual fiddling with negatives.
• With wrap off, it clamps: max(0, min(index, size - 1)). Index 12 into size 5 becomes 4; index -3 becomes 0.

Inputs:

• index (INT, default 0) - the value to normalize.
• size (INT, default 1) - the count, not the max index.
• wrap (BOOLEAN, default true) - wrap vs. clamp.

Output: index (INT).

Where it earns its place

The canonical use is looping. Drive a frame counter forward forever and route it through this node with size set to your batch or list length, and the output cycles 0, 1, 2, ..., n-1, 0, 1, ... instead of running off the end. Feed that into the pack's Switch Any (By Index) and you've got per-frame option switching with no bounds checking anywhere else. Clamp mode is the safer default when you'd rather hold the last valid value than cycle - say, for a progress-style selector that shouldn't jump back to the start.

It also pairs naturally with StringListStepper: the stepper already handles its own wrap internally, but Index Wrap is what you reach for when you're doing the indexing yourself - e.g., turning a raw count into a valid latent-batch index.

The one thing that trips people

size is the length of your collection, not the highest valid index. Five items means size = 5, and wrapped outputs are 0–4. Pass 5 thinking "max index" and index 5 wraps to 0 instead of clamping to 4 - a silent bug that's genuinely annoying to catch. When in doubt, count the items, not the slots.

Installation

Part of Vantage Nodes. In ComfyUI Manager search "Vantage Nodes," or:

cd ComfyUI/custom_nodes
git clone https://github.com/vantagewithai/Vantage-Nodes.git
pip install -r requirements.txt

Restart after install. Like the rest of the pack's math/control nodes, this one has no extra runtime dependencies beyond the base install - the heavy requirements.txt is for the GGUF and Qwen TTS pieces.

Troubleshooting

• Output never leaves a small range: that's wrap mode working as intended. Switch wrap off if you wanted clamping instead.
• Unexpected 0: check size - if it's 0 or negative the node returns 0 by design.