FC CropBottom

FC_CropBottom is an artifact of ComfyUI-FaceChain's node-naming history. Depending on where you're looking, "FC CropBottom" can mean two slightly different nodes: the currently-shipped one (registered as FC CropBottom, which takes a width and crops the bottom of the image to normalize dimensions) and this older/underscore variant, registered as FC_CropBottom, whose schema is built around a face_index input instead.

Both display as "FC CropBottom" in the UI, which is exactly the kind of confusion that wastes an afternoon, so let's get it straight. If you're writing a new workflow, use the current FC CropBottom node. If you've loaded an older workflow (or a tutorial that predates the pack's refactor) and ComfyUI tells you a node is missing, this is the name you're chasing.

The schema you'll see

Per the older node's definition:

• source_image (IMAGE) - the photo to process.
• face_index (INT, default 0) - an index into the detected faces.

One IMAGE output. In this older design, the crop was keyed to a face in the image rather than to a target width - a face-aware bottom crop, which the current implementation folded into the broader FC FaceDetectCrop / width-based flow. The current node replaced the face-index concept with a plain width parameter, which is simpler and easier to reason about.

What to actually do with it

• New graph: use FC CropBottom (the current node) - width input, resize-and-trim behavior.
• Old workflow: if it references FC_CropBottom, replace it with the current FC CropBottom or rebuild the crop step with FC FaceDetectCrop + a resize node. The face-index crop behavior mostly maps onto face_index in FC FaceDetectCrop's normal mode.

Install

Both names live in the same pack: ComfyUI Manager (search "ComfyUI-FaceChain") or git clone https://github.com/THtianhao/ComfyUI-FaceChain into custom_nodes/, then restart. The pack's startup dependency install (onnxruntime-gpu, mmdet, mmcv, modelscope, insightface, transformers…) applies either way, and any face-index-based crop also pulls in the ModelScope face-detection model on first use.

The short version: don't build anything new on this name. It's a legacy registration, kept alive in listings and old workflow JSONs, and the current FC CropBottom is the one you want.