An interactive visual editor for creating RGB color masks directly inside ComfyUI — with drag-and-drop shape manipulation, rotation, layers, and multiple shape types.

Features:

Interactive canvas editor with click-drag shape creation
9 shape types: rectangle, ellipse, triangle, diamond, pentagon, hexagon, star, arrow, cross
Move, resize, and rotate shapes with intuitive handles
Layer system with z-index ordering
Per-shape color, opacity, rotation controls
Undo/Redo (Ctrl+Z / Ctrl+Y)
Keyboard shortcuts (V = Select, R = Rect, E = Ellipse, Del = Delete)
Shift + rotate for 15° snapping
Resolution-independent (coordinates stored as percentages)
Live preview on the node after saving

RGB Mask Splitter 🔀

Companion node that splits an RGB mask image into separate masks by color matching with configurable tolerance. Useful for applying different conditioning/prompts to each color region.

Region Select ✂️

Pick one or more movable rectangular regions on an image and send each one out as a crop — ideal for a "crop a region → regenerate it → paste it back" (inpaint/variation) workflow. Each region carries its own prompt, and everything needed to reassemble the image travels in a single region output.

Features:

Interactive preview inside the node: drag each region box to position it
Flexible image source: connect an image_override, or load a file in the node
Up to 8 regions, added/removed with the + / − buttons (num_regions)
Shared region size (width / height); each box is positioned independently
One prompt box that edits the selected region (click a box to select it)
Outputs: crop + crop_2… (one per region), the full image (for compare), x / y / width / height, per-region prompt / prompt_2…, and a region bundle (carries the original image plus every region's coordinates and prompt)

Region Stitch 🧩

Companion node that pastes externally-generated variants back into their original positions. Connect the region bundle from Region Select plus each generated variant, and it composites them onto the original image — no manual coordinates needed.

Features:

Positions variants automatically from the region bundle (matched by number: variant → region 1, variant_2 → region 2, …)
Variant inputs appear progressively as you connect them
Feathered blending (blur) to hide seams; variants are auto-resized to their region
Optional per-region mask to paste only part of a variant (white = keep variant) — pairs nicely with Impact Pack's Preview Bridge for painting masks on the variant

Conditioning Mask Combine (dynamic) 🎭

Combines multiple (positive, negative, mask) triplets into a single positive/negative conditioning pair — a drop-in replacement for chains of native Conditioning (Set Mask) + Conditioning (Combine) nodes.

Features:

Dynamic triplets: the node starts with one slot and automatically reveals the next one each time you connect an input (up to 10)
Per-triplet mask_strength and set_cond_area (default / mask bounds) — no shared settings
Incomplete triplets are skipped silently, so unused slots never block the workflow
Pairs naturally with RGB Mask Splitter for regional prompting

Installation

Manual

cd ComfyUI/custom_nodes/
git clone https://github.com/Phykrex/ComfyUI-Momo-Nodes.git

Restart ComfyUI. No additional dependencies required (uses Pillow, torch, numpy — already included with ComfyUI).

ComfyUI Manager

Search for Momo Nodes in the ComfyUI Manager and install.

Usage

Find all nodes under the MomoNodes/ category in the right-click node menu.

RGB Mask Editor workflow

RGB Mask Editor ✏️ → RGB Mask Splitter 🔀 → Conditioning Mask Combine (dynamic) 🎭 → KSampler

Add RGB Mask Editor node, set width/height/background color
Click 🎨 Open RGB Mask Editor to launch the visual editor
Draw shapes with different colors for different regions
Apply & Close to save
Connect to RGB Mask Splitter to extract individual masks per color
Feed each mask together with its own positive/negative conditioning into Conditioning Mask Combine (dynamic) — new triplets appear as you connect
Pipe the combined positive/negative into your KSampler

Region variation workflow

Region Select ✂️ → (your sampler / edit model) → Region Stitch 🧩 → Save Image
                 └──────────────── region ──────────────┘

Add Region Select, load or connect an image, set width / height and num_regions
Drag each region box; click one to edit its prompt
Send each crop (+ prompt) through your own sampler to get a variant
Feed the region bundle and each variant into Region Stitch to reassemble the image

License

MIT