ComfyUI_StarNodes

ComfyUI_StarNodes

Little Helper Nodes For ComfyUI

StarNodes Theme System

You can now choose ready-to-use color themes for nodes in ComfyUI settings and apply theme presets via a node right-click menu (also supports multi-select). See: STARNODES THEME SYSTEM.md

Current Version: 2.3.7

<img width="917" alt="image" src="https://github.com/user-attachments/assets/4bc1378e-d1cf-4063-9196-b056a58444ec" />

A collection of utility nodes designed to simplify and enhance your ComfyUI workflows.

🎉 Version 2.3.7 - Star Lucida RMBG

🆕 New Node:

⭐ Star Lucida RMBG

Advanced background removal powered by Lucida (BiRefNet fine-tune). Excels at semi-transparent objects (glass), camouflage, text/logos with soft shadows, glow/VFX, and illustrations/line art. Auto-downloads model from HuggingFace on first use. Based on egeorcun/lucida.

---

🎉 Version 2.3.6 - New Nodes, Fancy Progress Bar & Consolidated Release

🆕 New Nodes:

⭐ Star LTXV All-in-One (2-Pass)

Single-node port of the "LTXV Sulphur All in 1" workflow: pass 1 at half resolution → 2x latent upscale → pass 2 at full resolution. Supports text-to-video, image-to-video, and image+audio-to-video modes with built-in model/LoRA/CLIP/VAE caching.

⭐ Star Video Loader

Standalone video loader that decodes any video file from the ComfyUI input folder into frames (IMAGE batch), audio (AUDIO), fps, frame count, and a STAR_FILENAMES reference — with force_rate / skip / every-kth / cap controls, inline preview and progress bars. No Video Helper Suite required.

⭐ Star Video Compressor

Compress videos or IMAGE batches to a chosen quality (CRF slider) or a desired file size (two-pass bitrate encoding). Formats: H.264 MP4, H.265/HEVC MP4, VP9 WebM, AV1 MP4. Perfect for Discord's 10 MB upload limit. Includes inline video preview, three progress indicators, and detailed info output. See web/docs/StarVideoCompressor.md.

⭐ Star Output Cleaner

Browse, select and clean up images in your ComfyUI output folder (or any custom folder) directly from a node. Features 200×200 thumbnails, paging, date filtering, permanent delete with confirmation dialog, and ZIP download.

⭐ Star Krea2 Unbound

Streamlined prompt adherence enhancer for Krea2 models. Single MODEL in, MODEL out — no widgets, no configuration.

⭐ Star 360 Parallax Viewer Pro

Enhanced panorama viewer with optimized frame rendering. Renders only one loop and duplicates frames for multi-loop exports. Includes "Create Video Frames" toggle, overlay effects, and depth map support.

⭐ Star Ollama Prompt Helper

User-friendly single node for local Ollama integration. Create or refine prompts using 15 preset system prompts or custom prompts. Supports vision models via optional image input.

⭐ Star Image Compare

Interactive before/after image comparison node with a draggable wipe slider.

⭐ Star Realistic Film Grain

Add realistic analog film grain with authentic film stock profiles, luminance masking, and organic variance.

⭐ Star Flux2/Qwen-Image-Edit Inpainter

All-in-one inpainting node for Flux2 and Qwen-Image-Edit models with built-in crop-and-stitch, reference image support, and Differential Diffusion.

⭐ Star Tiled SeedVR Upscaler

Upscales an image with SeedVR2 by processing it in overlapping tiles to keep VRAM usage low. Includes color correction for seamless blending.

⭐ Star Box Drawer

Draw rectangular boxes on images (filled or outlined) with customizable colors, positions, and sizes.

⭐ Star Image Shifter

Shift images horizontally and vertically with seamless wrapping for panoramas and tileable textures.

⭐ Star 360 Parallax Viewer

Interactive 360° panorama viewer with mouse-driven parallax, auto-rotation, fullscreen, and on-screen controls. Supports mono, SBS, and top/bottom stereoscopic images.

✨ Enhancements:

• Fancy animated DOM progress bar — Added to Star Sampler, Star SD Upscale Refiner (both variants), Star Tiled SeedVR Upscaler, Star 360 Parallax Viewer Pro, and Star LTXV All-in-One. The standard ComfyUI progress bar remains as fallback.
• Star Sampler (Unified) — Added tiled VAE decoder for lower VRAM usage during sampling
• Category Reorganization — All nodes organized under the ⭐StarNodes/ prefix
• Star Save Panorama JPG+ — Now outputs both original and stereoscopic 3D image

---

Version 2.0.1 - Major Release

This was a major release with new integrated nodes and cleanup of deprecated features.

New Integrated Nodes:

LTX Video Toolz (8 nodes)

• ⭐ Star LTX Video Settings - Comprehensive video dimension and frame calculator for LTX video generation with divisibility constraints
• ⭐ Star VAE LTXV Save - Advanced VAE encoder for LTX video with quality presets and latent saving
• ⭐ Star VAE LTXV Load - VAE decoder for LTX video latents
• ⭐ Star LTX Image Cut - Smart image cropping tool for LTX video with aspect ratio preservation
• ⭐ Star Multi Inputs to One - Combine multiple dynamic inputs into a single output
• ⭐ Star LTXV Get Last Frame - Extract the last frame from LTX video latents
• ⭐ Star LTXV Load Last Image - Load and process the last generated image
• ⭐ Star Video Joiner - Join multiple video files into a single seamless video

Music Generation (1 node)

• ⭐ Star ACE Step Music Generator (Local API) - Professional music generation using ACE Step 1.5 API with full control over duration, BPM, key/scale, lyrics in 50+ languages, and commercial-grade output in MP3/WAV/FLAC formats

⚠️ Breaking Changes - Removed Features:

• ❌ InfiniteYou Face Swap Suite (4 nodes) - Removed due to insightface dependency issues
• ❌ StarFaceLoader - Deprecated and unused
• ❌ StarGeminiRefiner - Removed to reduce external API dependencies
• ❌ StarFlowmatchOption - Deprecated experimental feature
• ❌ StarNanoBanana - Removed to reduce external API dependencies

What Changed:

• 86 active nodes (77 from cleanup + 9 new integrated nodes)
• Added soundfile>=0.12.0 dependency for music generation
• Removed google-generativeai and onnxruntime-gpu dependencies
• Cleaned up insightface-related commented dependencies
• Fixed duplicate imports bug
• Streamlined codebase for better maintainability
• Integrated LTX Video Toolz and ACE Step Music nodes

Total Active Nodes (v2.1.1)

• 89 active nodes (86 from v2.0.1 + 3 new helper nodes)

Migration: If you used any removed nodes, please check CHANGELOG.md for alternatives and migration notes.

---

New in 1.9.9

Dependencies

• 🔧 InfiniteYou Dependencies Now Optional — insightface, onnxruntime, huggingface_hub, and facexlib are now commented out in requirements.txt. Only install these if you need the InfiniteYou face swap nodes. This resolves numpy binary incompatibility issues for users who don't need face swap functionality.

Conditioning

• ⭐ Star Flux2 Conditioner (StarFlux2Conditioner) — Advanced conditioning node for Flux2 models that encodes text prompts and optionally processes up to 5 reference images. Automatically resizes images to 1MP and encodes them as guiding latents for enhanced image-to-image workflows.

Helpers & Tools

• ⭐ Star Size Calculator by Side (Star_Size_Calculator_By_Side) — Calculate new image dimensions by resizing the longest or shortest side while maintaining aspect ratio. Supports both image input and manual width/height. Outputs include width, height, long_side, and short_side (all as both string and int) for flexible workflow integration.

New in 1.9.7

Text & Data

• ⭐ Star Prompt Picker (StarPromptPicker) — Pick prompts from a text file (one prompt per line) or from a folder of single-prompt .txt files. Supports Random and One By One modes with saved progress.

Sampling Utilities

• ⭐ Distilled Optimizer (QWEN/ZIT) (StarDistilledOptimizerZIT) — Two-pass distilled refinement options for ⭐ StarSampler (Unified). Works with Z-Image-Turbo and with Qwen-Image when using a Turbo/LightX LoRA.

New in 1.9.5

IO & Metadata

• ⭐ Star Save Image+ (StarSaveImagePlus) — Save images with built-in folder/filename settings and store 5 extra metadata strings (StarMetaData 1-5) into the PNG.
• ⭐ Star Load Image+ (StarLoadImagePlus) — Load images and read out the 5 extra metadata strings (StarMetaData 1-5) as separate outputs.

Appearance

• ⭐ New Node Appearance Options — Right click a node to change its background and title bar colors.
• ⭐ Settings Option — In ComfyUI settings you can enable StarNodes style for every node: Apply StarNodes Style to All Nodes.
• Note: At the moment this only works in Nodes 1.0.

Workflows

• Info: Old workflows might need to be updated.
• ⭐ There is already a new template workflow for High Quality Z-Image-Turbo.

New in 1.9.4

Video & Animation

• ⭐ Star Image Loop (StarImageLoop) — Creates seamless looping video frames from images like panoramic images. Supports multiple dynamic image inputs that are joined horizontally to create longer slidess. Perfect for social media content from AI-generated or photographed panoramas.
• ⭐ Star Video Loop (StarVideoLoop) — Creates seamless looping video frames from video inputs. Videos are scrolled horizontally to create a slidingeffect with moving content. Supports multiple dynamic video inputs.

New in 1.9.3

Metadata & Workflow Sharing

• ⭐ Star Meta Injector (StarMetaInjector) — Transfers all PNG metadata (including ComfyUI workflow data) from a source image to a target image and saves it directly. Perfect for sharing workflows with custom preview images.

New in 1.9.2

Workflow Control & Preview

• ⭐ Star Stop And Go (StarStopAndGo) — Interactive workflow control node that lets you pause, preview, and decide whether to continue or stop your workflow. Works with any data type and supports user-select, timed pause, and bypass modes.

Model Tools & Conversion

• ⭐ Star Model Packer (StarModelPacker) — Combines split safetensors model files into a single file and converts them to a chosen floating-point precision (FP8, FP16, or FP32).
• ⭐ Star FP8 Converter (StarFP8Converter) — Converts existing .safetensors checkpoints to FP8 (float8_e4m3fn) and writes them into the standard ComfyUI output models folder.

New in 1.9.1

Image & Latent Utilities

• ⭐ Star Latent Resize (StarLatentResize) — Resizes existing latents to a target resolution using an advanced ratio/megapixel selector, with a custom mode for exact width/height while keeping model-friendly dimensions.

New in 1.9.0

Image Filters & Effects

• ⭐ Star Realistic Film Grain (StarRealisticFilmGrain) — Highly realistic analog film grain with authentic film stock profiles (Kodak Tri-X, Ilford HP5, Portra 400, etc.), luminance masking, organic variance, and soft-light blending for natural film aesthetics.
• ⭐ Star HighPass Filter (StarHighPassFilter) — High-pass based sharpening filter to enhance fine details and edge contrast.
• ⭐ Star Black And White (StarBlackAndWhite) — Flexible black-and-white conversion with tonal control for cinematic monochrome looks.
• ⭐ Star Radial Blur (StarRadialBlur) — Radial blur effect for focus/zoom style motion and creative depth effects.
• ⭐ Star Simple Filters (StarSimpleFilters) — Comprehensive image adjustment suite (sharpen, blur, saturation, contrast, brightness, temperature, color matching).

Workflow & Ratio Utilities

• ⭐ Star PSD Saver Adv. Layers (StarPSDSaverAdvLayers) — Advanced PSD exporter with enhanced layer handling for complex Photoshop workflows.
• ⭐ Star Advanced Ratio/Latent (StarAdvancedRatioLatent) — Combined advanced aspect ratio and latent megapixel helper for precise, resolution-safe size selection.

LoRA Utilities

• ⭐ Star Dynamic LoRA (StarDynamicLoRA) — Dynamic LoRA loader that lets you configure multiple LoRAs with flexible weights and options in a single node.
• ⭐ Star Dynamic LoRA (Model Only) (StarDynamicLoRAModelOnly) — Variant of Star Dynamic LoRA that only applies LoRAs to the model (no CLIP changes), ideal for more controlled style mixing.

Sampling Utilities

• ⭐ Distilled Optimizer (QWEN/ZIT) (StarDistilledOptimizerZIT) — Two-pass distilled refinement options for ⭐ StarSampler (Unified). Works with Z-Image-Turbo and with Qwen-Image when using a Turbo/LightX LoRA.

New in 1.8.0

Upscaling & Refinement

• ⭐ Star SD Upscale Refiner (StarSDUpscaleRefiner) — All-in-one SD1.5 upscaling and refinement node combining checkpoint loading, LoRA support, upscale models, tiled diffusion, ControlNet tile, and advanced optimizations (FreeU, PAG, Automatic CFG) into a single workflow.

LoRA Utilities

• ⭐ Star Random Lora Loader (StarRandomLoraLoader) — Randomly selects a LoRA from your library (with subfolder and name filters) and can optionally apply it directly to MODEL/CLIP or output the LoRA path as a string.

Previous Updates

1.7.0 – Qwen/WAN Image Editing & Utilities

Qwen/WAN Image Editing Suite

• ⭐ Star Qwen Image Ratio (StarQwenImageRatio) — Aspect ratio selector for Qwen models with SD3-optimized dimensions
• ⭐ Star Qwen / WAN Ratio (StarQwenWanRatio) — Unified ratio selector for Qwen and WAN video models with auto aspect ratio matching
• ⭐ Star Qwen Image Edit Inputs (StarQwenImageEditInputs) — Multi-image stitcher for Qwen editing (up to 4 images)
• ⭐ Star Qwen Edit Encoder (StarQwenEditEncoder) — Advanced CLIP encoder optimized for Qwen image editing
• ⭐ Star Image Edit for Qwen/Kontext (StarImageEditQwenKontext) — Dynamic prompt builder with customizable templates
• ⭐ Star Qwen Edit Plus Conditioner (StarQwenEditPlusConditioner) — Enhanced conditioning for Qwen models
• ⭐ Star Qwen Rebalance Prompter (StarQwenRebalancePrompter) — Intelligent prompt rebalancing for better results
• ⭐ Star Qwen Regional Prompter (StarQwenRegionalPrompter) — Region-based prompting for precise control

Image Processing & Effects

• ⭐ Star Apply Overlay (Depth) (StarApplyOverlayDepth) — Blend filtered images using depth/mask with Gaussian blur
• ⭐ Star Simple Filters (StarSimpleFilters) — Comprehensive image adjustments with color matching (sharpen, blur, saturation, etc.)

AI Generation & Prompting

• ⭐ Star Nano Banana (Gemini) (StarNanoBanana) — Google Gemini 2.5 Flash image generation with 30+ templates
• ⭐ Star Ollama Sysprompter (JC) (StarOllamaSysprompterJC) — Structured prompt builder for Ollama with art styles
• ⭐ Star Sampler (StarSampler) — Advanced sampler with extensive configuration options

Utilities & Tools

• ⭐ Star Save Folder String (StarSaveFolderString) — Flexible path builder with date-based organization
• ⭐ Star Duplicate Model Finder (StarDuplicateModelFinder) — SHA256-based duplicate model scanner

1.6.0 – IO & Image Utilities

• ⭐ Star Random Image Loader — Load random images from folders with seed control
• ⭐ Star Image Loader 1by1 — Sequential image loading with state persistence
• ⭐ Star Save Panorama JPEG — Export JPEGs with XMP panorama metadata
• ⭐ Star Frame From Video — Extract specific frames from video batches
• ⭐ Star Icon Exporter — Multi-size PNG/ICO export with effects

Available Nodes

⭐StarNodes/Starters

• ⭐ SD(XL) Starter: Loads checkpoint with CLIP and VAE, creates empty latent with customizable resolution
• ⭐ FLUX Starter: Loads Unet with 2 CLIPs and creates empty latent
• ⭐ SD3.0/3.5 Starter: Loads Unet with 3 CLIPs and creates empty latent

⭐StarNodes/Sampler

• ⭐ StarSampler SD/SDXL: Advanced sampler for SD, SDXL, SD3.5 with model and conditioning passthroughs
• ⭐ StarSampler FLUX: Specialized sampler for Flux models with model and conditioning passthroughs
• ⭐ Detail Star Daemon: Enhances image details, compatible with Flux and all SD Models (Adapted from original sources)
• ⭐ Star FluxFill Inpainter: Specialized inpainting node for Flux models with optimized conditioning and noise mask handling
• ⭐ Star Flux2/Qwen-Image-Edit Inpainter: All-in-one inpainting node for Flux2 (Dev/Klein) and Qwen-Image-Edit models with built-in crop-and-stitch, reference image support and Differential Diffusion
• ⭐ Star 3 LoRAs: Applies up to three LoRAs simultaneously to a model with individual weight controls for each

⭐StarNodes/Qwen & Image Editing

• ⭐ Star Qwen Image Ratio: Dropdown aspect ratio selector for Qwen models with SD3-optimized dimensions (1:1, 16:9, 9:16, 4:3, 3:4, etc.)
• ⭐ Star Qwen / WAN Ratio: Unified ratio selector supporting both Qwen and WAN video models with automatic aspect ratio matching
• ⭐ Star Qwen Image Edit Inputs: Prepares up to 4 input images for Qwen editing by intelligently stitching them into a single canvas
• ⭐ Star Qwen Edit Encoder: Advanced CLIP text encoder optimized for Qwen image editing with reference latents and caching
• ⭐ Star Image Edit for Qwen/Kontext: Dynamic prompt builder loading customizable templates from editprompts.json
• ⭐ Star Qwen Edit Plus Conditioner: Enhanced conditioning specifically designed for Qwen models
• ⭐ Star Qwen Rebalance Prompter: Intelligently rebalances prompts for optimal Qwen model performance
• ⭐ Star Qwen Regional Prompter: Region-based prompting system for precise control over different image areas
• ⭐ Star Apply Overlay (Depth): Blends filtered images over source using depth/mask with Gaussian blur options
• ⭐ Star Simple Filters: Comprehensive image adjustments (sharpen, blur, saturation, contrast, brightness, temperature) with advanced color matching
• ⭐ Star Nano Banana (Gemini): Google Gemini 2.5 Flash image generation/editing with 30+ prompt templates and flexible aspect ratios
• ⭐ Star Sampler: Advanced sampler with extensive configuration options for various workflows

⭐StarNodes/Image And Latent

• ⭐ Star Adaptive Detail Enhancer: Adaptively sharpens, denoises, and enhances image details using edge, face, and texture analysis. Great for portraits, art, and upscaling. See StarDetailEnhancer.md.
• ⭐ Star Image Compare: Compare two images with an interactive wipe slider. Far left shows Image 2, far right shows Image 1. See StarImageCompare.md.
• ⭐ Star Seven Inputs(img): Switch that automatically passes the first provided input image to the output
• ⭐ Star Seven Inputs(latent): Switch that automatically passes the first provided latent to the output
• ⭐ Star Face Loader: Specialized node for handling face-related operations. Image loader that works like the "load image" node but saves images in a special faces-folder for later use
• ⭐ Star Grid Composer: Compose multiple images into a grid layout with automatic sizing, captions, and customizable fonts/colors. Supports batch image and caption input via StarGridBatchers
• ⭐ Star Grid Image Batcher: Batch multiple images or image batches for use with Star Grid Composer, supporting up to 16 images
• ⭐ Star Grid Captions Batcher: Batch up to 16 caption strings for grid layouts in Star Grid Composer
• ⭐ Star Model Latent Upscaler: Complete pipeline for latent upscaling with model choice and VAE encoding/decoding
• ⭐ Star SD Upscale Refiner: All-in-one SD1.5 upscaling and refinement node with LoRA, upscale model, tiled diffusion, and ControlNet tile integration
• ⭐ StarWatermark: Adds customizable watermarks to images. Supports text, image, and advanced placement options for protecting or branding your outputs
• ⭐ Star 7 Layers 2 PSD: Saves up to seven images as layers in a single PSD file with automatic sizing based on the largest image dimensions
• ⭐ Starnodes Aspect Ratio Advanced: Enhanced version with additional options for aspect ratio calculation and resolution determination

⭐StarNodes/Text And Data

• ⭐ Star Seven Inputs(txt): Text concatenation with optional inputs. Works as automatic switch and concatenates multiple inputs
• ⭐ Star Text Filter: Cleans string text by removing text between two given words (default), removing text before a specific word, removing text after a specific word, removing empty lines, removing all whitespace, or stripping whitespace from line edges
• ⭐ Star Seven Wildcards: Advanced prompt maker with 7 inputs supporting wildcards and multiple random selections
• ⭐ Star Wildcards Advanced: Enhanced wildcard processing with support for folder paths, random selection, and multiple prompt inputs
• ⭐ Star Easy-Text-Storage: Save, load, and manage text snippets for reuse across workflows. Perfect for storing prompts, system messages, and other text content
• ⭐ Star Web Scraper (Headlines): Scrapes news headlines from websites for use in prompts or text generation
• ⭐ Star Ollama Sysprompter (JC): Builds structured prompts for Ollama with multiple art styles loaded from styles.json

⭐StarNodes/Video

• ⭐ Star Image Loop: Creates seamless looping video frames from panoramic images with dynamic multi-image input support
• ⭐ Star Video Loop: Creates seamless looping video frames from video inputs with dynamic multi-video input support

⭐StarNodes/IO

• ⭐ Star Meta Injector: Transfers PNG metadata (workflow, prompts, parameters) from source to target image and saves directly
• ⭐ Star Save Folder String: Flexible path builder for organized file saving with preset folders, date-based organization, and custom naming
• ⭐ Star Duplicate Model Finder: Scans ComfyUI models directory for duplicate files using SHA256 hashing with detailed reports
• ⭐ Star Random Image Loader: Load random images from folders with optional subfolders and seed control
• ⭐ Star Image Loader 1by1: Sequentially loads images across runs with state saved in the folder
• ⭐ Star Save Panorama JPEG: Save JPEGs with embedded XMP panorama metadata for 360° viewers
• ⭐ Star Frame From Video: Pick first/last/specific frame from an image batch (e.g., video)
• ⭐ Star Icon Exporter: Export multi-size PNGs and ICO with shaping, stroke, and shadow options
• ⭐ Star Save Image+: Save images like ComfyUI Save Image, but with 5 extra text inputs stored as StarMetaData 1-5 in the PNG metadata
• ⭐ Star Load Image+: Load images like ComfyUI Load Image, but with 5 extra text outputs read from StarMetaData 1-5 in the image metadata

⭐StarNodes/InfiniteYou

• ⭐ Star InfiniteYou Apply: Apply face identity from a reference image to generated images
• ⭐ Star InfiniteYou Face Swap Mod: Modified version of the face swap node with additional control options
• ⭐ Star InfiniteYou Patch Saver: Save face identity data for later use
• ⭐ Star InfiniteYou Patch Loader: Load previously saved face identity data
• ⭐ Star InfiniteYou Patch Combine: Combine multiple face patches with weighted influence
• ⭐ Star InfiniteYou Advanced Patch Maker: Create advanced face patches with detailed control options

⭐StarNodes/Conditioning

• ⭐ Star Conditioning I/O: Allows saving and loading conditioning information for reuse across workflows

⭐StarNodes/Settings

• ⭐ Star Save Sampler Settings: Save customizable sampling settings for StarSamplers with support for both SD and Flux samplers
• ⭐ Star Load Sampler Settings: Load previously saved sampling settings for StarSamplers
• ⭐ Star Delete Sampler Settings: Delete saved sampling settings

⭐StarNodes/Helpers And Tools

• ⭐ Star Size Calculator by Side: Calculate new dimensions by resizing longest or shortest side while maintaining aspect ratio. Outputs width, height, long_side, and short_side (both string and int)
• ⭐ Star Denoise Slider: Provides a simple slider interface to control the denoising strength for samplers
• ⭐ Starnodes Aspect Ratio: Calculates aspect ratio from an image or provides standard aspect ratios with customizable megapixel settings
• ⭐ Star Divisible Dimension: Ensures image dimensions are divisible by a specific value (useful for VAE compatibility)
• ⭐ Starnodes Aspect Video Ratio: Select a video aspect ratio from a dropdown, input width, and receive width/height as int/string plus formatted size (e.g., 750x422). Calculates height automatically from width and selected ratio.
• ⭐ Star Random Lora Loader: Randomly selects a LoRA from your library with subfolder/name filters and optional direct application to MODEL/CLIP or string output

⭐StarNodes/Color

• ⭐ Star Palette Extractor: Extracts dominant color palette from an image with various color format options

⭐StarNodes

• ⭐ Ollama Helper: Loads Ollama models from ollamamodels.txt for integration with Ollama nodes

*Note: You can add custom resolutions by editing the .json files in the node folder.

Documentation

Detailed documentation for all nodes is available in the web/docs directory of this repository. Each node has its own markdown file with comprehensive information about:

• Inputs and outputs
• Usage instructions
• Features and capabilities
• Technical details
• Tips and notes

The documentation is automatically loaded by ComfyUI when you access the help for any node, based on your locale settings.

Additional Documentation

• QwenEditPromptGuide.md - Comprehensive guide for using Qwen image editing nodes
• README_StarQwenRegionalPrompter.md - Detailed documentation for regional prompting
• SIMPLIFIED_REGIONAL_PROMPTER_V2.md - Simplified guide for regional prompter v2
• editprompts.json - Customizable prompt templates for Qwen/Kontext nodes
• styles.json - Art style definitions for Ollama Sysprompter

Installation

Via ComfyUI Manager (Recommended)

Search for "Starnodes" in ComfyUI Manager and install

Manual Installation

1. Open CMD within your custom nodes folder
2. Run: git clone https://github.com/Starnodes2024/ComfyUI_StarNodes
3. Restart ComfyUI

Find the nodes under "⭐StarNodes" category or search for "star" in the node browser.

Wildcards

You will find the wildcards in the wildcards folder of your ComfyUI main folder. If you add your own just copy the new files to this location.

Wildcard Rules in the Star Wildcards Node

Basic Wildcard Syntax

• Wildcards are defined using double underscores: __wildcard_name__
• The node looks for text files in the wildcards folder of your ComfyUI installation
• When a wildcard is encountered, a random line from the corresponding text file is selected

Folder Structure

• Wildcards can be organized in subfolders
• To use a wildcard in a subfolder, use the syntax: folder\__wildcard_name__
• The system will look for the file at [ComfyUI base path]/wildcards/folder/wildcard_name.txt

Random Options

• You can use curly braces {} with pipe symbols | to choose randomly between options
• Example: {option1|option2|option3} will randomly select one of the options
• You can even include wildcards inside these options: {__wildcard1__|__wildcard2__}

Nested Wildcards

• Wildcards can be nested within other wildcards
• The system supports up to 10 levels of recursion to prevent infinite loops
• When a wildcard contains another wildcard, the nested wildcard is also processed

Seed Behavior

• The node takes a seed parameter that determines the randomization
• Each prompt input (1-7) uses a different seed offset to ensure variety
• Each wildcard within a prompt also gets a unique seed based on its position

Error Handling

• If a wildcard file doesn't exist, the wildcard name itself is used as text
• If a wildcard file exists but is empty, the wildcard name is used as fallback
• If there's an error processing a wildcard, it falls back to using the wildcard name

The Star Wildcards node allows you to combine up to 7 different prompts, each with their own wildcards, which are then joined together with spaces to create the final output string.

• Supports recursive wildcard processing up to 10 layers

Requirements

Core Dependencies

All dependencies are listed in requirements.txt and will be installed automatically:

• color-matcher - Required for Star Simple Filters color matching features
• psd-tools>=1.10.0 - Required for PSD export nodes
• beautifulsoup4, newspaper3k, lxml - Required for web scraping nodes
• imageio-ffmpeg - Required for Star Video Loader/Compressor (bundled ffmpeg with libx264/libx265)