ComfyUI Custom Extensions

Specialized ComfyUI node designed to automate the creation of high-precision captioning datasets using multimodal LLMs. (Description by CC)

Tools about metadata.

Single-node image quality gate for LoRA dataset filtering. Analyzes blur, sharpness, noise, clipping, and entropy with three-tier verdict (PASS/SO-SO/FAIL), center-weighted blur detection, and shot-type presets.

Custom nodes for Qwen-Image-Edit with multi-image support, more flexibility around the vision transformer (qwen2.5-vl), custom system prompts, and some other experimental things to come.

Rembg Background Removal Node for ComfyUI

Pair prompts with their LoRAs on the canvas, then manage the rest from one panel — LoRAs, checkpoints, workflows and wildcards, with Civitai and Hugging Face search, server-side downloads, update checks, and a three-way merge that upgrades a workflow without losing your edits.

A ComfyUI custom node for integrating LLM API calls into workflows

Neo Image Chooser — a node for selecting images with a pleasant UI that allows pausing workflows for frame selection or image save confirmation. (Description by CC)

Random anime/game character picker with franchise pool management for ComfyUI.

A ComfyUI custom node for saving images in date directories like SD-Web-UI does.

Structured prompt-library and LoRA workflow nodes for ComfyUI that keep reusable prompt material outside workflow graphs.

A collection of utility nodes for ComfyUI to streamline workflows, including prompt editing, translation, model preset management, image preprocessing, and metadata editing.

Nodes for ComfyUI, mostly for llm calling and automatizing pulid multiface.

ComfyUI custom nodes in a single package: Arena legacy + AutoCache + Updater

NODES: Image Loader, Get Num Frames, Get Resolution Node, Video Loader, Film Grain

Multi-segment minimaxH3DYTsc 6.0 — official MiniMaxH3ImageToVideo / ReferenceToVideo pipeline.

Draw a mask at any point in a ComfyUI workflow (after resize / rotate / crop), keep every mask as a numbered file, and shape the mask edge.

ComfyUI custom nodes for Nucleus-Image 17B MoE diffusion model with FP8 and bf16 precision support, dynamic block swap, CFG rescaling, and model shift parameters. (Description by CC)

Shamelessly copied the code to auto color correct the image like in gimp from this answer: [a/https://stackoverflow.com/a/56365560/4561887](https://stackoverflow.com/a/56365560/4561887)

A collection of A5 prompt, text, image, and latent nodes for ComfyUI.

Pause ComfyUI workflows to inspect and manually select images from a batch.

A lightweight node to compress and preview images using the temp directory, displaying file sizes on the node UI.

ComfyUI nodes for video -> audio -> JSON transcript pipeline. Designed to pair with ComfyUI-MiMoASR for the ASR step.

A ComfyUI custom node for inpainting workflows with integrated Photopea editing and adjustable region selection. Features: adjustable reference region, integrated Photopea image editing, professional mask editing with feathering control, auto mask detection, multi-language support.

Classifier-free guidance taken apart into seven stage nodes (when, weak branch, combine, where, correct, govern, measure) plus 50 nodes named after the published guidance methods they implement (APG, CFG-Zero*, PAG, SEG, Perp-Neg, FDG, the guidance interval and more), all on one engine. SDXL, SD1.5 and flow-matching models (Anima).

ComfyUI custom node for browsing Civitai image prompts directly inside a node.

A suite of ComfyUI nodes for generating parameter sweep video matrices with labeled grids. Create comprehensive visual comparisons of how different parameter combinations affect your video generation results.

Lightweight ComfyUI node for text-prompted segmentation masks ◧ using CLIPSeg and mask auto-aligned trim. numpy 2.x safe, TensorFlow-free, auto model caching.

ComfyUI custom node pack with tagger functionality combining WD14, Camie, and PixAI models. (Description by CC)

Companion extension node for comfyui-adaptiveprompts that adds advanced, experimental, or specific workflow-related features.

Generates audio from music prompt using MusicGen

Converts image description to music prompt

Scan ComfyUI workflows for missing model metadata and download models into registered model folders with resumable transfers.

comfyui-zeug (German for 'gear' or 'stuff') is a collection of custom nodes for ComfyUI, designed to enhance functionality and provide additional features.

A ComfyUI node to manipulate 3D objects.

Custom nodes for the Foundation-1 audio model in ComfyUI, providing BPM-aware generation and streamlined audio workflows.

A collection of utility nodes for ComfyUI to improve workflow efficiency.

ComfyUI节点 - 将多张图像拼接成网格图。支持动态输入（1-32张图像）、多种分辨率预设、行列设置，自动为每个网格添加顺序编号。

When your workflow starts acting weird and you need to know what's actually happening to your data - not just guess from looking at the output.

Experimental MiniMax H3 frame-rate adjustment with temporal-coordinate retiming and pre-sampling AV duration alignment for ComfyUI.

A ComfyUI custom node that refines and quantizes messy AI-generated pixel art into clean, perfect, and grid-aligned pixels.

ComfyUI custom node for offline Chinese speech-to-text using FunASR with VAD and punctuation support.

Extract keyframes from video

Extended HTTP request nodes for ComfyUI — supports image, audio, video binary data, form-data upload, and raw binary I/O. Fork of felixszeto/ComfyUI-RequestNodes.

A ComfyUI plugin for uploading generated images and videos to Alibaba Cloud OSS (Object Storage Service).

Fail-safe Google Translate prompt node for ComfyUI (retry + caching).

Node for compositing multiple images with interactive preview and layer management

This repository contains custom nodes for Stability AI API which supports SD3.0 and 3.5.

ComfyUI ChatGPT Helper ComfyUI ChatGPT Helper is a custom node extension for ComfyUI that integrates OpenAI's ChatGPT capabilities directly into your ComfyUI workflows. This tool allows for dynamic prompt generation, automated text manipulation, and enhanced interactivity within your AI image generation processes.

FeiFei's custom nodes: LLM-driven Prompt Director, prompt enhancing, vision captioner, aspect-ratio sizing, style/character templates, watermark, WebP save/load.

Nodes: LiamLoadImage. This node provides the capability to load images from a URL.

Comprehensive QwenVL integration for ComfyUI with HuggingFace and GGUF model support

A lightweight ComfyUI custom node for LoRA weight optimization using CMA-ES and Tournament GA algorithms.

A ComfyUI custom node pack for calling LLM providers.

This is an audio processing script that applies soft limiting, optional loudness normalization, and optional slicing for transcription. It can also produce stereo-mixed outputs with optional audio appended to the end. The script organizes processed files into structured folders with sanitized filenames and retains original timestamps for continuity.

Convert a structured scene description into an optimised SDXL prompt via LLM

ComfyUI custom node for building varied prompts across batch generations using sections and per-section pick modes.

Simple and efficient custom node for overlaying customizable text onto images with font support, multi-language ready, and full position/color control.

ComfyUI nodes for creating images with the look and feel of various film stocks.

Transforms a string by dynamically evaluating Python code written in the node

A creative alternative to dynamicprompts

An optimized, pure PyTorch implementation of the Ideogram 4 logit-normal timestep scheduler for ComfyUI.

ComfyUI UNO Nodes is a collection of nodes for ComfyUI that allows you to load and use UNO models.

ComfyUI custom nodes integrating three Chinese AI image generation platforms (Hi Dream, Douban, NanoBanana) supporting text-to-image and image-to-image generation with parallel processing and robust error handling. (Description by CC)

A comprehensive, all-in-one suite for seamlessly integrating Large Language Models (LLMs) like Google Gemini and OpenAI (ChatGPT) directly into your ComfyUI workflows.

multiGPU batch-parallel upscaling nodes for ComfyUI.

This repository contains a custom node for ComfyUI that converts text into a JSON object. The node is designed to be user-friendly and supports multi-line JSON input.

NanoBanana2 is a custom ComfyUI node for Google Gemini image generation models.

Inline help panel for ComfyUI Nodes with remote HTML/Markdown support.

Extension widgets for ComfyUI

Tile-based upscaling nodes for ComfyUI — Nano Banana 2 compatible

ComfyUI custom nodes wrapping the Lightricks LTX REST API (text-to-video, image-to-video, retake/video-to-video, and HDR EXR video-to-video).

A custom ComfyUI node that randomizes tag weights for text prompts. Useful for dynamic prompt generation or creative experimentation in AI image workflows.

ComfyUI nodes for 3D visualization.

ComfyUI node package for BYOK (bring-your-own-key) access to external generative APIs: Gemini, Veo, and more via LiteLLM.

Enhance GLB 3D models with realistic materials, smart emissive elements, and procedural normal maps

This is a pack with some handy nodes for ComfyUI.

Text-to-motion animation pipeline for ComfyUI: SOTA motion models (MDM, priorMDM, MoMask, Kimodo) in isolated Docker containers, plus resampling, IK cleanup, character rigging, and GLB/FBX export nodes.

ComfyUI custom nodes for ByteDance's Lance—unified 3B multimodal model supporting text/image/video generation, editing, and understanding in single framework.

Node: utils/Simple Counter This node is a simple counter, when pressing 'Queue Prompt' resets the count.

load and download YouTube videos from URL's for quick and easy use for vid2vid cases.

Kokoro is an open-weight TTS model with 82 million parameters.

Interpolate sdxl latents using slerp with and without a mask. use with unsample nodes for best effect

SDXL model loader supporting custom clip_g and clip_l embeddings with NVFP4 optimization. (Description by CC)

This extension provides PSD layer saving functionalities with mask support for ComfyUI.

ComfyUI nodes for Qwen3-TTS: voice design, voice library, segmented generation

A collection of efficient workflow nodes for ComfyUI with advanced sampling, seed management, temporal prompt processing, and workflow optimization tools

1 node to have control over your workflow.

Utility nodes for ComfyUI: Prompt Stack, Picker, Search & Replace, and Labeled Index.

ComfyUI node for working with complex prompts with support for variables, randomness, and basic logic.

ComfyUI nodes for detailed region inpainting with smart cropping, upscaling, and seamless compositing back into the original image.

ComfyUI node to make memes and demotivators

ComfyUI prompt enhance nodes and string utilities

Utility ComfyUI custom nodes that add some extra features for your workflows (random int for cli runs, unix timestamp, boolean random, string match switch, etc).

Resolution picker for low-base + upscale workflows: preset sizes, custom ratio or exact manual W×H, two upscale stages (base → mid → final), plus an empty LATENT and a live readout. Built for Krea 2 / Selfism.

Backend liaison services for SugarSubstitute and ComfyUI.

32-aligned image preprocessing node for MiniMax H3 workflows

A python port of pixelit by giventofly.

Qwen2 Nodes for ComfyUI. I needed to run Qwen2 on ComfyUI to use it in my workflow for batching images and captioning and none of the implementations I found on the web worked the way I wanted.[w/May contain bugs.]