凤希AI - 图片列表合并

FxAiImageBatchConcat does one thing: it takes two image lists, 图片列表A and 图片列表B, and returns them concatenated as 合并图片 plus a 总数量 count. That's the whole node. If that sounds too trivial to write about, remember that in segmented-video workflows - this pack's entire reason for existing - you're constantly assembling lists out of pieces: the tail of one segment plus the head of the next, frame batches from two loaders, reference images from two asset folders. A node that just chains them without complaining saves you a lot of fiddly rewiring.

How it works

The mechanism is short enough to quote from the source: it normalizes each input (a native ComfyUI batch tensor becomes a Python list of frames, and a list stays a list), extends list A with list B, and returns the combined list plus its length.

Two details matter in practice. First, it handles both ComfyUI's standard [B, H, W, C] batch tensor and the pack's own Python-list-of-frames format - the pack's image manager and batch-load nodes can hand you either, and this node accepts both, which is most of its value. Second, it's pure append: nothing gets sorted, deduplicated, or reordered. Whatever order A then B is, that's the output order, which is exactly what you want for frame sequences where order is meaning.

Inputs and outputs

• 图片列表A, 图片列表B - the two IMAGE inputs. Both are required on the schema, though the code quietly tolerates a None on either side and just concatenates what's there.
• 合并图片 - the appended result, same format you fed in.
• 总数量 - len(A) + len(B). Handy for wiring into a loop counter or frame-count check so a downstream node knows what it's dealing with.

Installing it

It's part of fxai-toolkit (author 凤希AI, MIT). Install once via ComfyUI Manager (search "fxai-toolkit") or:

cd ComfyUI/custom_nodes
git clone https://github.com/fxai666/fxai-toolkit
# restart ComfyUI

No models, no dependencies beyond the pack's usual first-load auto-install of soundfile and psutil. Chinese-only labels, support via QQ group 775649071 and Bilibili.

Where people get burned

Mostly by expecting it to do more. It won't resize mismatched frames to a common resolution - if A is 512×512 and B is 960×544, the merged list is a mess of mixed sizes and you'll need to run the result through a uniform-size node (this pack has one) before a sampler touches it. It also won't dedupe, so concatenating the same list twice doubles your frames. And it's not a "merge to a grid" node - that's a different utility (the batch-load node here produces grid/contact-sheet output separately). For plain concatenation it's the right tool, and nothing else to it.