ComfyAngel

Custom nodes for ComfyUI that make batch workflows easy and visual editing intuitive.

• Text Permutation — Generate prompt combinations from {cat,dog} on {chair,sofa} syntax
• Loop Nodes — True loop without Auto Queue, process each item and collect results
• Parameter Overlay — Display generation metadata on images for sharing and comparison
• Visual Widgets — Position picker, crop editor, color picker with interactive UI

🎬 Video Tutorial:

Screenshots

Text Permutation — Easy to create text combo list to batch change prompt

Load Images Batch from Folder — Load all images with validation and history

Load Image + Metadata | Parameter Overlay | Image Bridge — Easy to place watermark

Loop Start | Loop End — Easy batch processing each item from items list (supports text and image)

Smart Composite — Visual Position Picker

Smart Crop — Visual Crop Editor

Widget Nodes — Color Picker, Resolution Picker, Solid Color

---

Node Index (A-Z)

Node	Description
Color Picker	Interactive color picker with eyedropper
Custom Text Overlay	Add custom text overlay with full styling
Image Bridge	Preview/save images without breaking the chain
Image Info	Get dimensions and batch info
JSON Extract	Extract values from JSON using field paths
Load Image + Metadata	Load image and extract generation metadata
Load Images Batch from Folder	Load images from folder with validation
Loop End	End loop and collect results
Loop Start	Start loop over items
Parameter Overlay	Add generation parameters as visual overlay
Parameter Parser	Parse metadata and extract parameters
Resolution Picker	Select from common resolutions by aspect ratio
Smart Composite	Composite overlay with visual position picker
Smart Crop	Crop image with visual editor
Solid Color	Generate solid color image
Split Image Batch	Extract single image from batch by index
Text Combine	Combine multiple text strings
Text Permutation	Generate all combinations from template
Workflow Metadata	Output current workflow as JSON

---

Installation

ComfyUI Registry (Recommended)

comfy node install comfyangel

ComfyUI Manager

Search for "ComfyAngel" in ComfyUI Manager and click Install.

Manual Installation

cd ComfyUI/custom_nodes
git clone https://github.com/ThepExcel/ComfyAngel.git

Then restart ComfyUI.

---

Nodes

Loop Nodes

True loop functionality without Auto Queue. Iterate through batches/lists and accumulate results in a single execution.

---

Loop Start 🪽

Start a loop over items.

Input	Type	Description
items	ANY	Batch/list to iterate over. Accepts IMAGE batch, lists of strings, or any iterable.

Output	Type	Description
flow	FLOW_CONTROL	Must connect to Loop End.
item	ANY	Current item from the batch for this iteration.

Example - Loop through text prompts:

[Text Permutation] outputs: ["cat on chair", "cat on sofa", "dog on chair", "dog on sofa"]
         ↓
[Loop Start] iteration 0: item = "cat on chair"
[Loop Start] iteration 1: item = "cat on sofa"
[Loop Start] iteration 2: item = "dog on chair"
[Loop Start] iteration 3: item = "dog on sofa"

---

Loop End 🪽

End a loop and collect all results into a single output.

Input	Type	Description
flow	FLOW_CONTROL	Must connect from Loop Start.
result1	ANY	Value to collect from each iteration (optional).
result2	ANY	Second value to collect (optional).

Output	Type	Description
results1	ANY	All accumulated results from result1.
results2	ANY	All accumulated results from result2.

v0.6.0: Loop End now executes even without downstream connections. You can run loops without connecting nodes to the results outputs.

Accumulation behavior:

• IMAGE/MASK tensors: Concatenated into a batch tensor. 4 images → tensor shape (4, H, W, C)
• Strings: Collected into a Python list.
• Other types: Collected into a Python list.

---

Loader Nodes

Load Images Batch from Folder 🪽

Load images from a folder with folder browser, validation, and history.

Input	Type	Description
folder_path	STRING	Path to folder. Use Validate Folder button to check.
sort_by	ENUM	name, modified_date, created_date
max_images	INT	Maximum images to load. 0 = unlimited.
start_index	INT	Skip first N images.
include_subdirs	BOOLEAN	Scan subfolders recursively.
resize_mode	ENUM	disabled, resize_to_first, resize_to_largest, resize_to_smallest

Output	Type	Description
images	IMAGE[]	List of images (preserves RGBA when present).
filenames	STRING	Newline-separated list of loaded filenames.
count	INT	Number of images loaded.

Features:

• Validate Folder button: Shows image count and validates path
• Folder history: Recent folders saved for quick access
• RGBA preservation: Images with alpha channels are preserved as 4-channel tensors

---

Split Image Batch 🪽

Extract a single image from a batch by index. Useful for Auto Queue workflows.

Input	Type	Description
images	IMAGE	Batch of images to split.
index	INT	Which image to extract (0-based). Auto-increments with Auto Queue.
loop	BOOLEAN	Wrap around when index exceeds batch size. Default: true

Output	Type	Description
image	IMAGE	Single image at the specified index.
index	INT	The actual index used.
total_count	INT	Total images in the batch.

---

Text Utility Nodes

Text Permutation 🪽

Generate all combinations from a template with inline options.

Input	Type	Description
template	STRING	Template with {option1,option2} syntax.
separator	STRING	Separator inside braces. Default: ,
trim_options	BOOLEAN	Remove whitespace around options. Default: true

Output	Type	Description
texts	STRING[]	List of all combinations.
count	INT	Number of combinations.

Example:

Template: "a {cat,dog} on a {chair,sofa}"

Output (4 combinations):
- "a cat on a chair"
- "a cat on a sofa"
- "a dog on a chair"
- "a dog on a sofa"

---

Text Combine 🪽

Combine multiple text strings into one.

Input	Type	Description
text1-4	STRING	Texts to combine (text1 required, others optional).
separator	STRING	String between texts. Default: \n

Output	Type	Description
combined	STRING	All non-empty texts joined.

---

JSON Extract 🪽

Extract values from JSON using field paths.

Input	Type	Description
json_string	STRING	JSON text to parse.
fields	STRING	Field paths, one per line.
delimiter	ENUM	Field separator: newline, comma, pipe, semicolon
output_delimiter	ENUM	Output value separator.
default_value	STRING	Value when field not found.

Output	Type	Description
values	STRING[]	List of extracted values.
values_joined	STRING	All values joined.
count	INT	Number of fields extracted.

Field path syntax:

seed                    → json["seed"]
3.inputs.steps          → json["3"]["inputs"]["steps"]
nodes[0].type           → json["nodes"][0]["type"]

Browse JSON Fields button: Opens visual JSON tree viewer.

---

Overlay Nodes

Load Image + Metadata 🪽

Load an image and extract embedded generation metadata.

Input	Type	Description
image	IMAGE_UPLOAD	Select image file.

Output	Type	Description
image	IMAGE	The loaded image.
mask	MASK	Alpha channel as mask.
metadata_raw	STRING	Raw metadata string.
metadata_formatted	STRING	Human-readable formatted parameters.

Supported formats: A1111/Civitai, ComfyUI workflow embedding

---

Parameter Parser 🪽

Parse metadata string and extract individual parameters.

Input	Type	Description
metadata	STRING	Raw metadata string.
show_prompt	BOOLEAN	Include prompt in output.
max_prompt_length	INT	Truncate prompt length.

Output	Type	Description
formatted	STRING	Human-readable summary.
model	STRING	Checkpoint/model name.
sampler	STRING	Sampler name.
seed	INT	Generation seed.
steps	INT	Sampling steps.
cfg	FLOAT	CFG scale.
positive_prompt	STRING	Positive prompt.
negative_prompt	STRING	Negative prompt.

---

Parameter Overlay 🪽

Add generation parameters as a visual overlay on an image.

Input	Type	Description
image	IMAGE	Image to add overlay to.
image_path	STRING	Optional. Path to read metadata from.
metadata_text	STRING	Optional. Raw metadata text.
position	ENUM	bottom_extend, bottom_inside, top_inside
font_size	INT	Text size 8-100. Default: 25
bg_opacity	FLOAT	Background opacity 0.0-1.0. Default: 0.7
show_prompt	BOOLEAN	Include prompt text. Default: false
max_prompt_length	INT	Truncate prompt length. Default: 200

Output	Type	Description
image	IMAGE	Image with parameter overlay.

---

Custom Text Overlay 🪽

Add custom text overlay with full styling control.

Input	Type	Description
image	IMAGE	Image to add overlay to.
text	STRING	Text to display (supports \n).
position	ENUM	bottom_extend, bottom_inside, top_inside
font_size	INT	Text size 8-100.
bg_opacity	FLOAT	Background opacity.
text_color	STRING	Text color hex. Default: #FFFFFF
bg_color	STRING	Background color hex. Default: #000000

Output	Type	Description
image	IMAGE	Image with text overlay.

---

Composite Nodes

Smart Composite 🪽

Composite (layer) an overlay image onto a canvas with visual position picker.

Input	Type	Description
canvas	IMAGE	Background image (connect via Image Bridge for preview).
overlay	IMAGE	Image to place on top.
quick_position	ENUM	Reference point: free, top_left, top_center, top_right, middle_left, center, middle_right, bottom_left, bottom_center, bottom_right
x	INT	Horizontal offset from quick_position.
y	INT	Vertical offset from quick_position.
anchor	ENUM	Which point of overlay aligns to position.
scale_percent	FLOAT	Resize overlay 1-500%. Default: 100
blend_mode	ENUM	normal, multiply, screen, overlay, soft_light, hard_light, difference, add, subtract
opacity	FLOAT	Overlay transparency 0-100%.
mask	MASK	Optional. External mask for overlay transparency.
invert_mask	BOOLEAN	Invert the mask (ComfyUI LoadImage masks are inverted). Default: true

Output	Type	Description
image	IMAGE	Composited result.
x	INT	Final X position.
y	INT	Final Y position.

v0.6.0 Updates:

• Quick Position as reference point: X,Y are now offsets from the quick position, not absolute coordinates
• Mask input: Connect MASK output from LoadImage for overlay transparency
• Invert mask option: ComfyUI's LoadImage MASK is inverted (0=opaque), enable invert_mask to correct this

Visual Editor Features:

• Drag overlay directly on preview
• Quick position buttons for corners/edges/center
• Scale slider with live preview

---

Utility Nodes

Image Bridge 🪽

Pass-through node that previews or saves images without breaking the chain.

Input	Type	Description
image	IMAGE	Image(s) to preview/save. Handles different sizes.
mode	ENUM	preview or save
filename_prefix	STRING	Prefix for saved files.

Output	Type	Description
image	IMAGE	Same image passed through.

Essential for loops: Place inside loop to see each iteration's result.

v0.6.0: Now handles batches with different-sized images correctly.

---

Workflow Metadata 🪽

Output current workflow as JSON strings.

Output	Type	Description
prompt_json	STRING	Current prompt as JSON.
workflow_json	STRING	Full workflow as JSON.

---

Image Info 🪽

Get dimensions and batch info from an image.

Input	Type	Description
image	IMAGE	Image tensor.

Output	Type	Description
width	INT	Image width.
height	INT	Image height.
channels	INT	Number of channels.
batch_size	INT	Number of images in batch.

---

Widget Nodes

Resolution Picker 🪽

Select from common image resolutions by aspect ratio.

Input	Type	Description
aspect_ratio	ENUM	1:1, 4:3, 3:4, 16:9, 9:16, 21:9, 3:2, 2:3
resolution	ENUM	Specific resolution (options change based on ratio).

Output	Type	Description
width	INT	Selected width.
height	INT	Selected height.
aspect_ratio	STRING	The ratio string.

Presets: SD 1.5, SDXL, SD3/Flux, DALL-E 3, Midjourney, Standard (FHD, 4K)

---

Solid Color 🪽

Generate a solid color image.

Input	Type	Description
color	STRING	Hex color (e.g., #FF0000).
width	INT	Image width.
height	INT	Image height.
batch_size	INT	Number of images.

Output	Type	Description
image	IMAGE	Solid color image.

---

Color Picker 🪽

Interactive color picker with eyedropper.

Input	Type	Description
color_hex	STRING	Color in hex format.
image	IMAGE	Optional. Connect for eyedropper.

Output	Type	Description
color	STRING	Normalized hex color.

Features: Visual picker, HEX/RGB/HSL input, color presets, eyedropper tool

---

Smart Crop 🪽

Crop image with visual editor.

Input	Type	Description
image	IMAGE	Image to crop.
x	INT	Left edge of crop.
y	INT	Top edge of crop.
crop_width	INT	Width of crop.
crop_height	INT	Height of crop.
bg_color	STRING	Background color for out-of-bounds areas.

Output	Type	Description
image	IMAGE	Cropped image.

Visual editor: Click button to open drag-to-select crop area.

---

Tips & Best Practices

Loop Performance

• Image Bridge is essential - Without it, you won't see intermediate results
• No downstream required - Loop End now runs even without connected outputs (v0.6.0)
• Memory usage - Each iteration accumulates in memory. For large batches, use Split Image Batch with Auto Queue

Smart Composite Workflow

[Load Image] → [Image Bridge] → canvas [Smart Composite]
                                overlay ← [Load Logo]
                                mask ← [Load Logo].mask (with invert_mask: true)

Metadata Workflow

[Load Image + Metadata] → metadata_raw
         ↓
[Parameter Parser] → formatted, model, seed, etc.
         ↓
[Parameter Overlay] ← original image
         ↓
[Save Image]

---

Example Workflows

Download and drag into ComfyUI to try:

Workflow	Description
Text Permutation	Generate prompt combinations and batch process
Load Images Batch from Folder	Load all images from a folder
LoadImageBatch + Loop	Load folder of images and process each with Loop nodes
Smart Crop	Visual crop editor demo
Color & Resolution Picker	Widget nodes for color and resolution selection

---

Changelog

v0.6.0

• Loop End: Now executes without downstream connections
• Smart Composite: Quick position is now a reference point (X,Y are offsets)
• Smart Composite: Added mask input with invert_mask option for transparency
• Image Bridge: Handles different-sized images in batches
• Load Images Batch: Preserves RGBA channels

v0.5.0

• Added JSON Extract node with visual field browser
• Image Bridge preview improvements
• Various bug fixes

---

Links

• GitHub: https://github.com/ThepExcel/ComfyAngel
• Registry: https://registry.comfy.org/thepexcel/comfyangel

---

License

MIT