ComfyUI Custom Extensions
Browse 9,446 ComfyUI extensions and 73,215 custom nodes.
A collection of custom nodes to consolidate functionality of many other nodes, enable unified workflows, and add custom functionality you won't find with other nodes.
Provides text randomization, concatenation, and wildcard loading utilities for ComfyUI workflows. (Description by CC)
A modular toolkit for advanced prompt engineering in ComfyUI designed for dataset creation, LoRA training, and controlled variation.
A generic, flexible node to integrate any WaveSpeed model into ComfyUI — including Nano Banana Pro, Wan 2.1, Flux, Kling, Veo, and more.
Nodes for ComfyUI in order to generate battelmaps
A ComfyUI custom node that exposes REST API endpoints for managing ComfyUI custom nodes and models.
Low-memory BF16 CUDA/CPU/SSD-offload ComfyUI nodes for HunyuanImage-3.0-Instruct
A collection of custom nodes for ComfyUI designed to facilitate complex looping and iteration workflows, specialized in media batch processing.
ComfyUI nodes for the Krea API: text-to-image, editing, masked edits, bounding-box region edits and enhancers across every Krea image model.
Pads an image to a square by extending its shorter side, with an optional longest-side resize, and crops the original image back out with the mask it outputs.
This node allow to intercept changes on the input string and choose between use the current one or sync with incoming new one.
This project contains custom Flip-Flop nodes for ComfyUI.
NODES:ImageMask2PNG
ComfyUI custom node package for tag-based prompt RAG using CSV datasets and local LLM generation.
Width and height from an input image's aspect ratio and a megapixel target, with a live output-size preview.
ComfyUI custom node for direct OpenAI image generation API calls using user-provided API key, with automatic text-to-image/image-to-image switching. (Description by CC)
A custom node for ComfyUI. Based on the specified resolution, it randomly selects and outputs one of three patterns: original orientation, swapped width/height, or a square format. (Description by CC)
Custom ComfyUI node that compresses generated files into ZIP archives for cloud environments like RunPod and Google Colab. (Description by CC)
Frame chaining for AI video in ComfyUI: extract a clip's last frame, feed it to the next image-to-video generation, and stitch the clips into one seamless continuous shot (with optional speed ramp).
Nodos custom, al estilo fury, nodos que añaden versatilidad. con el tiempo ira creciendo los nodos.
Professional AI art direction and advanced workflow suite for ComfyUI.
node set for image harmonization and composition
ComfyUI custom nodes for LiteTracker, a fast and efficient point tracking model for videos with load model, track, and grid editor nodes.
Extract authentic noise patterns from real photographs and apply them to AI-generated or composited elements for seamless photorealistic integration.
ComfyUI custom nodes for Nano Banana (Gemini) image generation
Create 2-4 characters with individual prompts
One-click group bypasser: a single node listing every group in your workflow as a button to enable or bypass it, with position, alphabetical or custom ordering.
Allows the locking of a nodes value from previous run
Preprocessor Contact Sheet for ComfyUI - runs all ControlNet preprocessors at once, displays them in a visual grid, and lets you inject AIO or dedicated preprocessor workflows with click-to-place.
A dedicated suite of custom nodes for batch testing artists, styles, and prompts in ComfyUI, solving the caching problem with anti-cache logic and dynamic filename generation.
A Single Node to replace WebUI workflow in ComfyUI.
A powerful ComfyUI custom node for model comparison testing supporting 1-5 checkpoint models with automatic VRAM management and comparison image generation. (Description by CC)
ComfyUI node for consuming and processing repository content
Custom node for integrating with gadzoinks iPhone app
Prompt engineering, Danbooru dictionary lookup, and AI prompt refinement nodes for ComfyUI.
Nodes that I needed but couldnt find, so ended up making.
A small ComfyUI custom node for shortening long silent sections in AUDIO inputs.
A ComfyUI custom node adapter for Xiaomi MiMo-V2.5-ASR.
A ComfyUI custom node to load SVG files and convert them to raster images, with options for resizing, background color, and borders.
ComfyUI custom nodes for YOLO detection and mask processing
Lightweight ComfyUI custom node pack to manage reusable sampler/scheduler presets with configuration and selection separated in two nodes.
A generation and curation workspace for ComfyUI: every generated image arrives with its parameters in one window, so you can sort them into persistent trays, compare models or crops side by side and upscale with one click, with no digging through output folders and image viewers. Exchanges layers and masks with Photoshop, Krita and GIMP.
ComfyUI custom-node pack and installer for hidden-image workflows across SD 1.5, SDXL, and FLUX with pattern preprocessing, text-mask generation, and workflow blueprints.
Production-ready ComfyUI port of ByteDance's LatentSync pipeline delivering high-quality lip synchronization, smooth mouth articulation, and VRAM-efficient inference for video content.
ComfyUI custom nodes for universal video import/export workflows
24oiduts ComfyUI
Offload image upscaling to a remote server with a dedicated GPU. Real-time progress streaming via SSE.
ComfyUI extension for FLUX.2 with SDNQ quantization.
Smart and practical custom nodes for ComfyUI.
A lightweight version of the custom nodes originally developed by [a/ManglerFTW](https://github.com/ManglerFTW/ComfyI2I) for performing image-to-image tasks in ComfyUI.
GS SmartSwitch - a boolean switch that automatically grays out widgets on whichever branch is inactive, including promoted parameters on the outer subgraph face, so dead calculation paths don't leave confusing editable fields lying around.
Sidebar panel for ComfyUI that auto-captures every prompt run, diffs the workflow graph between any two runs (added/removed/type-swapped/changed nodes, including inside subgraphs), locates changes directly on the canvas, and pairs the diff with a drag-and-drop wipe-compare of the outputs.
Generates dynamic combo/dropdown choices from multiline text inputs compatible with Set/Get node setups.
Custom ComfyUI node for image upload, upscaling, VAE encoding, and empty latent creation with dynamic scaling support. (Description by CC)
Custom nodes package organized by functional domains with namespace support for image upscaling, utilities, image repair, inpainting, and professional editing workflows. (Description by CC)
Gather statistics and information about your workflow.
Nodes:Change Image Border
Nodes:Apply Kmeans Filter
Nodes:Apply Meanshift Filter
A ComfyUI custom node that names saved outputs after the selected upscale model.
Welcome to the Image Upscaler & Restorer project! This tool utilizes the [a/GFPGAN](https://github.com/TencentARC/GFPGAN) algorithm to enhance and restore images, providing a seamless way to improve image quality.
Generate multiple groups of seeds for batch image processing in ComfyUI.
ComfyUI-SendToDiscord is a custom node for ComfyUI that simplifies sending preview images to Discord via webhooks. It supports both single-image uploads and batch mode, making it an efficient tool for sharing your generated images directly with your Discord server.
Nodes:NSFW Detector
Download missing models from the current ComfyUI workflow with live progress.
Just a simple substring node that takes text and length as input, and outputs the first length characters.
API plugin for Alibaba Cloud Wanx 2.1 Image Editing model. Supports 10 functions including stylization, inpainting, and outpainting.
ComfyUI custom nodes for automatic watermark masking and direct removal with OpenCV, EasyOCR, and Big-LaMa.
ComfyUI custom node for image generation with Alibaba Cloud DashScope.
ComfyUI custom nodes for OpenAI's GPT Image 2 with text-to-image, image editing, and cost estimation capabilities.
ComfyUI node providing Tetris game functionality. (Description by CC)
These nodes allow your YouTube LiveStream viewers to create on your local ComfyUI, you can make this a members only feature with a screen behind you displaying your members creations
Finishing nodes that make generated images read more like photographs: lens vignette and colour fringes, film halation, and grain that behaves like film.
WarpPipe - A data bundling and transfer system for ComfyUI workflows
An all-in-one ComfyUI node suite that simplifies model management, streamlines workflows, and makes prompting easier.
BPM-synced sparse track point generator and BPM detector for ComfyUI
A custom node that allows opening file paths in the default system file explorer.
Custom nodes for running NVIDIA Maxine NIMs from ComfyUI.
ElevenLabs API调用节点
One-click export workflow as API JSON (with LoadImage placeholder replacement), PNG image, or both bundled as ZIP.
Kimodo text-to-motion nodes and production bridges for Mixamo, Unity Humanoid, and Rive 2D characters
Green-screen cleanup, chroma keying, despill, transparent GIF/WebP export, and stable dynamic square cropping for ComfyUI.
Beat-synced flashback video generator with optical flow morphing — warp between original/edited image pairs in sync with music beats
MiniMax cloud video generation node. Prompt + optional first-frame image in, video out. Inference runs on MiniMax cloud, no local model loaded.
A ComfyUI plugin to manage local AI models and generated images through a dedicated web interface
A ComfyUI custom node that queues one job per prompt. Dynamically add or remove prompt fields as needed.
SHIFT reference-image style steering for FLUX.2 Klein (4B, 4B Base, 9B, 9B Base): extract an activation-steering vector from one style image, then generate plain text-to-image in that style, with no reference image at inference.
ComfyUI custom node pack for embedding workflow data in image pixels using steganography
Extension nodes to help bridge StableProjectorz and ComfyUI
CineStack H3: local cinematic workflows with fixed per-segment seeds and strict cached second-pass refinement.
12 workflows de ComfyUI convertidos en una casa encantada: cada uno con su panel, su personaje, su música y su voz · 12 ComfyUI workflows turned into a haunted house, each with its own panel, character, music and voice.
12 workflows de ComfyUI en un solo panel cyberpunk: imagen, edición, vídeo, interpolación y música, sin tocar nodos · 12 ComfyUI workflows in one cyberpunk panel: image, editing, video, interpolation and music, no node wiring.
These are designed to be custom nodes i found usefulness to that hopefully others can share.
ComfyUI Video Auto-Cropping Tool — includes keyframe detection and automatic segmentation/saving based on keyframes.
ComfyUI custom convenience nodes: Cropping images to a given aspect ratio, Cropping images to a max/min aspect ratio, If/else logic with provided float (rather than using a float node)
LAN discovery and shared compute, local Topaz Video AI processing, LLM workflows, and named exports for ComfyUI.
ComfyUI text input node with prefix/suffix blocks and Python-slice-style block selection