IncrementBatch

Let's be honest about IncrementBatch: the name promises arithmetic, and the code delivers a round trip. In the current TBG ETUR build, the entire implementation is run(cls, seed): return (seed,). You feed it an INT, it hands the same INT back. It does not increment anything.

So why does it exist? In ComfyUI, the thing that "increments" is the seed widget's control after generate behavior - you set a seed, run, and the widget advances. That works great inside a single sampler node, but it doesn't give you a wire you can branch, log, or feed into a second node that needs the same seed. IncrementBatch is a carrier: it takes the seed and exposes it as a real INT output you can route around the graph. In ETUR's batch workflows, where one seed needs to reach the tiler, the refiner, and a couple of helper nodes without being re-rolled each run, that passthrough is the point.

What it actually does

One input, one output:

• seed (INT) - the value you want to carry. Default 4, range up to the full 64-bit seed space.
• INT (output) - the same value, now available on a wire.

That's the whole schema. If you were hoping for a batch counter that spits out 4, 5, 6, 7 per batch item - not in this version. If you genuinely need an incrementing sequence, the stock ComfyUI Seed node with "increment" selected under control-after-generate, or a small expression node, will do it more honestly. Use IncrementBatch when you want the current seed captured and shared, not when you want a counter.

Installing it

Same pack, same drill:

cd ComfyUI/custom_nodes
git clone https://github.com/Ltamann/ComfyUI-TBG-ETUR
cd ComfyUI-TBG-ETUR
pip install -r requirements.txt

Or ComfyUI Manager → search "TBG" → install TBG_Enhanced Tiled Upscaler & Refiner FLUX PRO, restart. Yes, you're pulling the whole heavy pack for a passthrough - this is a helper node that came along for the ride, not a reason to install TBG.

Notes

The honest take: don't build a workflow around this node's name. Treat it as a seed-routing utility and you'll never be confused by it. If you load an older ETUR workflow that chains these and expect it to behave like a batch iterator, you'll get the same seed on every output - which is sometimes exactly what you want (consistent seeds across tiles) and sometimes not. In ETUR's own tiled workflows, consistent seeds across a batch are a feature, not a bug: the whole Neuro-Generative Tile Fusion design relies on neighboring tiles agreeing on their noise rather than diverging. So in context, the passthrough isn't a bug either. It's just a node whose job description changed and the label never caught up.