Production-focused custom nodes for ComfyUI workflows: image loading, tiling, compositing, mask utilities, smart resize, logic routing, prompt/text tooling, batch helpers, indexed model loaders, garment/portrait segmentation, and VTON preprocessing.

Author: jetthuang.com
Hugging Face: nhathoangfoto
Support: PayPal

Why NH-Nodes

NH-Nodes is built for practical workflow construction. It focuses on reusable building blocks that remove repetitive graph work:

Load images with filename/path metadata.
Split images into tiles, upscale or process tiles, then stitch them back.
Build image grids, layer composites, comparisons, and labeled previews.
Resize images and masks with predictable aspect-ratio behavior.
Route workflow branches with boolean logic and generic data switches.
Generate prompts from templates, regex, lists, schedules, and text utilities.
Pick indexed LoRA/diffusion models for larger configurable workflows.
Prepare VTON masks, agnostic images, DensePose/parsing outputs, and related masks.

Installation

ComfyUI Manager

Search for NH-Nodes, install, then restart ComfyUI.

Comfy CLI

comfy node install nh-nodes

Manual

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

Restart ComfyUI after installation.

Featured Workflows

Tile, Process, Untile

Use this when an image is too large for a model, or when every tile needs the same enhancement/upscale step.

Load Image -> Image Tile (NH) -> upscale/process tiles -> Image Untile (NH)

Image Tile (NH) supports:

original_ratio: derive tile dimensions from the original image aspect ratio.
custom: use explicit tile width and height.
square: force square tiles, for example 1024 x 1024.
Automatic overlap calculation.
Edge padding so square/custom tile outputs keep the requested tile size.

Image Untile (NH) uses tile_data from the tile node to stitch results back. It also handles tiles that were resized or upscaled, such as 2x or 3x per tile.

Grid Contact Sheet

Image batch -> Image Grid Composite (NH) -> Preview Image

Use manual cell size or first-image cell size. Resize modes include stretch, pad, and fill.

Metadata-Aware Image Load

Load Image Info (NH) -> image / mask / file_name / path

Useful for naming outputs, logging source images, or routing by filename.

Smart Resolution Resize

Load Image -> NH Smart Ratio Image Resize -> (your pipeline) -> NH Remove Resize Padding

NH Smart Ratio Image Resize picks the closest model preset itself from model_family + resolution_level, so it does not need NH Smart Resolution Picker upstream. Resize with cover crop, contain pad, or stretch behavior.

In contain_pad mode it also emits resize_data describing the content region, padding, and scale factor. Feed that into NH Remove Resize Padding at the end of the pipeline to crop the padded border back off (content_region) or to restore the original input size (original_size).

For cover_crop, use NH Restore Image Geometry after processing. Connect the processed image to image, the source image to original_image, and the resize node's resize_data to the matching input. It returns an image at the exact original dimensions, with processed content placed back in its original region and cropped-away borders retained from the source. Its mask is white in the restored region and black in the retained borders; use it for masked blending with the original, optionally softening the mask edges first.

Processed images may be upscaled, including by a non-integer factor, provided their framing is unchanged. Additional crops, rotations, or model-generated changes in geometry cannot be reversed by this metadata. The restore node also accepts contain_pad and stretch metadata and supports matching batches or singleton batch broadcasting. Restoration uses bicubic (default), bilinear, or nearest sampling; it does not recover details lost during downsampling.

Use NH Smart Resolution Picker when you need target dimensions and a latent for text-to-image, instead of resizing an existing image.

Garment and Portrait Segmentation

Load Image -> Garment Segment (NH) [part=upper_garment] -> mask / cutout_rgba / mask_image
Load Image -> Portrait Segment (NH) [hair + face] -> mask / cutout_rgba / mask_image

SegFormer-based masks at input resolution. cutout_rgba is the input image with the mask as alpha (RGB is left untouched outside the mask); mask_image is the mask as an RGB preview. Dresses and jumpsuits belong to upper_garment. Both nodes share mask_expand, mask_blur, and fill_holes. Choose part=all_garments to combine upper/lower garments, footwear and headwear in one mask, for example a shirt and pants together. attach_parts still controls whether touching pockets, zippers and other garment decorations are included. Models are auto-downloaded to models/NH-Nodes/ on first use (existing models/face_parsing or models/RMBG/segformer_fashion copies are reused).

Node Catalog

Image Loading and Saving

Node	Purpose
`Load Image Info (NH)`	Load one image and output `IMAGE`, `MASK`, file name, and path.
`Save Image Path (NH)`	Save image batches to a chosen folder and output saved filenames/path/count.
`NH Save Large Image`	Save full-resolution images without attaching heavy full-res previews to the canvas.
`Load Images Folder (NH)`	Load images from a folder by index/step, with filename and path outputs.
`Load images matching`	Load source/matching images by text, folder context, and matching rules.

Image Tiling, Grid, and Composition

Node	Purpose
`Image Tile (NH)`	Split image batches into overlapping tiles with metadata for reconstruction.
`Image Untile (NH)`	Rebuild full images from processed/upscaled tiles and tile metadata.
`Image Grid Composite (NH)`	Combine an image batch into a configurable grid.
`NH Large Image Preview`	Preview large images with lightweight canvas thumbnails and passthrough output.
`NH Large Image Compare`	Compare large A/B images with lightweight canvas proxy and tiled slider/side/difference viewer.
`NH Large Preview Cache Manager`	Inspect and clean the NH large preview tile cache by age, size, or cache id.
`Layer Layout Composite (NH)`	Place layers into preset, grid, or custom layout slots.
`Layer Stack Composite (NH)`	Paste multiple layers onto a background using explicit coordinates.
`Image Resize Unit (NH)`	Resize by pixel/mm/cm using stretch, pad, crop, or lock-ratio modes.
`Image Compare (NH)`	Compare two images side-by-side, top/bottom, split, difference, or overlay.
`Image Label (NH)`	Add header/bottom labels with font, padding, color, and outline controls.
`NH Agnostic Image Generator`	Create agnostic/masked/composite images for inpaint and VTON workflows.
`NH Simple Face Paste`	Paste a source face region into a target using mask feathering.

Smart Resolution

Node	Purpose
`NH Smart Resolution Picker`	Pick model/preset-aware dimensions and latent setup.
`NH Smart Ratio Image Resize`	Auto-pick the closest model preset for the source aspect, then resize with cover, contain, or stretch; outputs `resize_data` for padding removal.
`NH Ratio Preset Image Resize`	Resize to a chosen model preset ratio with crop, pad, or fill behavior.
`NH Remove Resize Padding`	Remove the padded border from processed images using resize metadata.
`NH Restore Image Geometry`	Restore processed or upscaled images to original coordinates with a blend mask; keep source pixels outside the crop.

Vision Segmentation

Node	Purpose
`Garment Segment (NH)`	Mask `upper_garment` (tops, jackets, coats, dresses, ties, scarves, collars, sleeves), `lower_garment` (pants, shorts, skirts, belts, tights; no shoes), `footwear`, `headwear`, or combine all four with `all_garments`. Attached parts (pockets, zippers, bows...) merge only when touching the selected garments.
`Portrait Segment (NH)`	Checklist mask of face, neck, hair, eyes, lips, nose, eyebrows, ears (+ optional mouth interior). `face` is the solid face region including features.

Mask Tools

Node	Purpose
`Mask Morphology (NH)`	Expand/shrink masks, fill holes, and blur edges.
`Mask Properties (NH)`	Output mask bounding box, dimensions, and related measurements.
`NH Create Box Mask`	Convert a mask area into a rectangular box mask.
`NH Mask Aspect Ratio Match`	Adjust one mask to match another mask's aspect ratio.
`NH Mask-Aware Resize`	Resize an image/mask pair while preserving the masked subject region.

Logic and Flow Control

Node	Purpose
`Compare (NH)`	Compare values with typed operators.
`Logic Gate (NH)`	AND, OR, NOT, XOR, NAND, and NOR boolean logic.
`If/Else (NH)`	Select between two inputs by condition.
`Switch N (NH)`	Pick one of several generic inputs by index.
`Any Switch (NH)`	Boolean switch helper.
`Any Branch Switch (NH)`	Route generic data through selected branches.
`Gate Switch (NH)`	Pass or block a generic input by boolean state.
`Value Match Index (NH)`	Return index of matching configured value.
`Math Eval (NH)`	Safe math expression evaluator with numeric/string outputs.
`Random Choice (NH)`	Weighted random choice from generic inputs.

Text and Prompt Tools

Node	Purpose
`String Operations (NH)`	Strip, case convert, replace, slice, inspect, and test strings.
`Prompt Join (NH)`	Join text fragments with separators and empty-line handling.
`Text Split (NH)`	Split text by delimiter and expose count/first/last values.
`Regex Extract (NH)`	Extract, replace, split, and match with regex patterns.
`Text Index (NH)`	Select a line/item from text by index.
`Text Concatenate (NH)`	Concatenate multiple text inputs.
`Text Split Lines (NH)`	Split multiline text into selected parts and metadata.
`Text Random Line (NH)`	Pick a random line from multiline text.
`Prompt Template (NH)`	Fill `{placeholder}` variables from connected text values.
`Prompt Scheduler (NH)`	Select prompts by step, sequence, ping-pong, or seed.

Lists and Batch Helpers

Node	Purpose
`List Create (NH)`	Build an `NH_LIST` from multiline or delimited text.
`List Index (NH)`	Select an item from an `NH_LIST`.
`List Filter (NH)`	Filter lists with contains, regex, equals, length, and related conditions.
`Batch Index (NH)`	Extract one or more images from an image batch.
`Batch Merge (NH)`	Merge image batches with resize handling.
`Counter (NH)`	Stateful counter for repeated queue runs.
`Any List Split (NH)`	Split any incoming list into an `All` output plus up to 10 indexed outputs.

Indexed Loaders

Node	Purpose
`Load LoRA Model Index (NH)`	Apply one selected LoRA to `MODEL` by 1-based index.
`Load LoRA Clip Index (NH)`	Apply one selected LoRA to `CLIP` by 1-based index.
`Load Diffusion Model Index (NH)`	Load one diffusion model by 1-based index.
`Load VAE Index (NH)`	Load one VAE by 1-based index or boolean toggle.
`Load Clip Index (NH)`	Load one CLIP by 1-based index or boolean toggle, with CLIP type selection.

Workflow Utilities

Node	Purpose
`Multi-Slider (FLOAT)`	Five float controls in one node.
`Multi-Slider (INT)`	Five integer controls in one node.
`Universal Slider Builder (NH)`	Generate dynamic slider outputs from text config.
`Boolean Switch (NH)`	Simple boolean toggle.
`Pack Universal (NH)`	Bundle generic values into an `NH_UNIVERSAL_PIPE`.
`Unpack Universal (NH)`	Extract values from an `NH_UNIVERSAL_PIPE`.

VTON Preprocessing

Node	Purpose
`VTON Ultimate Processor (NH)`	Generate masks, agnostic image data, DensePose/parsing outputs, hair/hands masks, and related VTON preprocessing assets.

NH-Nodes/
  Image/
    Load/save helpers
    Tile/untile
    Grid and layer composition
    Resize, compare, label
  Mask/
    Morphology, properties, bbox, aspect matching
  Vision/
    Garment and portrait segmentation
  Logic/
    Compare, gates, switches, branching, math, random choice
  Text/
    String, regex, split, concatenate, templates, scheduling
  Batch/
    Image batch, list helpers, any-list split
  Resolution/
    Smart resolution picker, ratio resize, preset resize, padding removal
  Loaders/
    Indexed LoRA, diffusion model, VAE, and CLIP loaders
  Utils/
    Sliders, booleans, universal pipe
  VTON/
    VTON Ultimate Processor

Requirements

numpy
Pillow
scipy
opencv-python
scikit-image>=0.18.0
transformers>=4.20.0
tqdm>=4.62.0
huggingface_hub>=0.16.0

VTON Ultimate Processor (NH) also needs ONNX Runtime, which is not installed automatically (the CPU wheel would overwrite an existing GPU build). Install one: python -m pip install onnxruntime-gpu (NVIDIA) or python -m pip install onnxruntime (CPU).

Most image/text/logic nodes use common ComfyUI dependencies. VTON and Vision nodes download model assets on first use (transformers SegFormer weights for Vision: ~190 MB fashion, ~340 MB face).

Development Notes

Layout: core/ holds shared helpers and model backends (no nodes); nodes/<category>/ mirrors the ComfyUI menu, and every module there that defines NODE_CLASS_MAPPINGS is registered automatically by __init__.py - drop a file in the right folder and restart. VTON's vendored preprocessing lives in nodes/vton/preprocess/.
Tests: python -m pytest custom_nodes/NH-Nodes/tests -q from the ComfyUI root.
Restart ComfyUI after installing or updating nodes.
If a node appears cached in the browser, refresh the ComfyUI page after restart.
Image Tile (NH) and Image Untile (NH) must pass tile_data directly through the workflow so reconstruction uses the original tile layout.

License

NH-Nodes' own code is MIT. VTON Ultimate Processor (NH) and Portrait Segment (NH) depend on third-party code or models that are licensed for non-commercial use only, and the model behind Garment Segment (NH) declares no license. See THIRD_PARTY_NOTICES.md before using these three nodes commercially.

If NH-Nodes helps your workflow, consider starring the repository: github.com/jetthuangai/NH-Nodes

README.md

NH-Nodes for ComfyUI