ComfyUI Custom Extensions
Browse 9,420 ComfyUI extensions and 73,215 custom nodes.
ComfyUI nodes for NVIDIA LocateAnything-3B visual grounding, detection, text localization, GUI grounding, and pointing.
Add-on nodes for ComfyUI-H3-FaceRefine (MiniMax H3 video face refiner, required). v0.2.0: GPU Flow Lock and Tone Match (2.7x faster pipeline), one Mode dropdown for V1 (prompt + 3 refs) or V2 (SAM3 face mask, optional VOSR2), default strength 0.30. Also Stitch Plus and super-resolved face crops. Side-by-side tests: https://annicodex.github.io/FaceRefine-Plus-Tests/
Nodes:Create Project Root, Add Folder, Add Folder Advanced, Add File Name Prefix, Add File Name Prefix Advanced, ShowPath
Portrait Tools: Facial detection cropping, alignment, ID photo, etc.
ComfyUI dust3r
ComfyUI-KikoTools provides carefully crafted, production-ready nodes grouped under the 'ComfyAssets' category. Each tool is designed with clean interfaces, comprehensive testing, and optimized performance for SDXL and FLUX workflows.
ComfyUI custom node that brings a DAW-style interactive video timeline directly into the node graph. Upload any video, scrub through it in real time, drag a loop region to define your extraction window, and pipe the resulting frame batch into any downstream node.
A fast and flexible random prompt generator with 12 columns (Empty / Pre-filled SFW / Pre-filled NSFW). Pre filled with text so that you can get started right away. Supports comma or newline output and always adds quality tags.
An AI movie studio inside ComfyUI: a browser-based studio with timeline, per-scene prompts and one-click generation, plus a self-improving engine that learns from your ratings. Every feature is a module, so a broken one is simply absent.
Krea2 / Krea2 Turbo style semantic conditioning and img2img fusion nodes for ComfyUI.
ComfyUI nodes for Genfocus generative refocusing. Features DeblurNet (restore all-in-focus images), BokehNet (create realistic depth-of-field/bokeh effects), DepthPro depth estimation, and defocus map utilities. Uses FLUX.1-dev with specialized LoRA adapters.
ComfyUI nodes for Qwen-Image-Edit and Qwen-Image — 30 nodes for editing, generation, multi-stage, LoRA, Spectrum acceleration, ControlNet-guided generation, prompt rewriting, and 2× upscale VAE
MiniMax H3 prompt timelines, beat-planned sampling, temporal reshots, native latent upscaling, pixel-space refinement, and shot assembly for ComfyUI
An unofficial Python library for [a/Suno AI](https://www.suno.ai/) API
SaltAI AudioViz contains ComfyUI nodes for generating complex audio reactive visualizations
This is an image recognition node for ComfyUI based on the RAM++ model from [a/xinyu1205](https://huggingface.co/xinyu1205). This node outputs a string of tags with all the recognized objects and elements in the image in English or Chinese language. For image tagging and captioning.
Adds seed-dependent noise to text conditioning to increase output diversity for Krea 2 Turbo (port of SeedVarianceEnhancer v2.2).
A custom node for ComfyUI that integrates Video-As-Prompt for motion-guided video generation from image inputs.
An awesome light image processing tool nodes for ComfyUI.
These are ComfyUI nodes to assist in converting an image to sketches or lineArts.
ComfyUI Frontend Vue Basic is custom node that demonstrate how to use vue as frontend framework along with primevue and vue-i18n, cooperating with ComfyUI API. Needs ComfyUI Frontend 1.25.0 or later.
Spectral analog film simulation (negative, enlarger print, scan) for ComfyUI. Film modeling powered by spektrafilm.
Unofficial ComfyUI implementation of [a/UniWorld-V1](https://github.com/PKU-YuanGroup/UniWorld-V1).
Zero-1-to-3: Zero-shot One Image to 3D Object, unofficial porting of original [Zero123](https://github.com/cvlab-columbia/zero123)
Give each reference image a job, and every attribute a dial. Guide cards assign per-image roles (keep the same subject, suggest the visual style, copy lighting or pose, big shapes only, avoid copying text/logos) with per-card strength, timing, and user-defined recipes - and Concept Slider cards turn any attribute you can name (brightness, age, fog density) into a training-free -6..+6 dial derived from Krea 2 text encoding.
These are companion nodes for Texturaizer, a Blender plugin that connects complex 3D data to ComfyUI. https://texturaizer.com
Advanced nodes of the Level Pixel company (levelpixel, LP). Includes convenient advanced nodes for working with LLM и VLM models (LLaVa) with GGUF format. Qwen2.5-VL and Qwen2.5 supported. Also included is a node for the RAM model. Nodes have the ability to automatically unload models from VRAM.
ComfyUI LC MaskMaker Nodes by lonecatone23: on-node Outpaint and Create Mask, background removal, Segment Anything, Smart Detailer (SAM 3, replaces SEGS / Impact detailers), person masks, mask refine, image blend, auto adjust, depth and normal maps. Models load on demand and are labelled with their licenses. Companion to ComfyUI_LC123_nodes; separate Registry name.
DWPose & NLF skeleton editor, retargeter, and renderer for ComfyUI. Extract body/hand/face keypoints (+ optional NLF 3D) from video, edit them frame-by-frame in a full-screen timeline + dope sheet + graph + 3D orbit editor with animated camera, retarget DWPose proportions per-cluster, and render clean pose images. Outputs standard ComfyUI types — drives WanAnimate, ControlNet, SCAIL, LTX-Video, UniAnimate, and any other pose-driven pipeline.
Solricks is an Istanbul-based technology brand focused on AI systems, custom software, cloud infrastructure, hardware, and enterprise technology solutions.
toobusy folds tedious multi-step ComfyUI workflows into single production nodes.
A simple ComfyUI extension to export images of your workflows with embedded JSON data.
A set of custom nodes designed for experiments with video diffusion models.
Nodes for Comfyui to use Prompt Quill within complex workflows
ComfyUI port of Vectorscope CC and Diffusion Color Grading by Haoming02. Makes it possible to adjust Brightness/Contrast/Saturation/Hue during image generation.
Nodes: AstropulsePixelDetector, BackgroundRemover, ImagePadForBetterOutpaint, InpaintingPipelineLoader, Inpainting, ...
Multi-LoRA loader with optional LTX layer-specific strength control, drag-and-drop file import, and a companion cycle node for batch-testing LoRA training checkpoints at multiple strengths. Outputs modified model plus JSON metadata of applied LoRAs.
RENAMED TO ComfyUI-MultiLoRALoader STILL HAS LTX STUFF AND EASY MIGRATION PATH
Custom ComfyUI nodes for loading and saving images to a PixlStash vault.
A dual-model image-to-video sampler for ComfyUI that combines high-noise and low-noise models in a single node, replacing two KSamplerAdvanced nodes with zero overhead and bit-identical frame generation. (Description by CC)
Latent space walks for latents, conditionals and noise
A ComfyUI custom node package that allows downloading and organizing Instagram content directly in your ComfyUI Output folder
Nodes: CLIPTextEncodeA1111, RerouteTextForCLIPTextEncodeA1111.
This node pack allows loading of SD checkpoints that uses diffusers format in comfyUI.
A ComfyUI custom node implementation for optimized video generation and motion effects, designed to work with Hunyuan text-to-video models.
A ComfyUI custom node that integrates Mistral AI's Pixtral Large vision model, enabling powerful multimodal AI capabilities within ComfyUI. Pixtral Large is a 124B parameter model (123B decoder + 1B visual encoder)
A ComfyUI wrapper implementation of ThinkSound - an advanced AI model for generating high-quality audio from text descriptions and video content using Chain-of-Thought (CoT) reasoning.
Transparent WebM overlay for live previews and sampler previews in ComfyUI.
Flux Pro via Replicate API Create API key at [a/https://replicate.com/account/api-tokens](https://replicate.com/account/api-tokens) Copy config.ini.example to config.ini and put the replicate key there.
Easy quant comfyUI origin model to gguf , and esay use it, save more disk...
Nodes: Chatbox Overlay. Custom node for ComfyUI to add a text box over a processed image before save node.
ComfyUI V3 nodes for FireRedAudio using an isolated Transformers 5.8 worker
Flow Upscaler is a fast latent upscaler model that works in the Flux.2 latent space, using a lightweight Rectified Flow model with 59M parameters for single-step upscaling.
A quick and easy ComfyUI custom node for setting SDXL-friendly aspect ratios.
HSWQ SDXL/Z Image ConvRot INT8/Hybrid ConvRot NVFP4 Loader, KREA2 ConvRot INT8 Loader and various tools
An advanced LoRA stacker node specifically designed for Nunchaku FLUX models, featuring dynamic UI control for easy and efficient LoRA management.
ComfyuiImageBlender is a custom node for ComfyUI. It may be used to blend two images together using a specified blending mode.
MuseTalk ComfyUI Preprocess and Postprocess Nodes
ComfyUI nodes for editing background of images/videos with CUDA acceleration support.
Runs the MoGe2 model on the input image. v1: Ruicheng/moge-vitl v2: Ruicheng/moge-2-vitl-normal
This custom node helps to auto download models from huggingface
a comfyui custom node for [a/I2V-Adapter](https://github.com/KwaiVGI/I2V-Adapter)
ComfyUI custom nodes for AI image generation using SynVow API supporting T2I (Text-to-Image) and I2I (Image-to-Image) with multiple API sources and batch generation. (Description by CC)
Seamless Clone for ComfyUI
ComfyUI nodes from [Timesaver](https://github.com/AlexYez/comfyui-timesaver).
This node is mainly based on the Yolov8 model for object detection, and it outputs related images, masks, and JSON information.
Nodes: CLIP Directional Prompt Attention Encode. Direction prompt attention tries to solve the problem of contextual words (or parts of the prompt) having an effect on much later or irrelevant parts of the prompt.
Prompt Rewriter for ComfyUI using llama.cpp for local inference. Supports LLMs, VLMs, and both Instruct and Thinking models.
Custom node for ComfyUI providing easy resolution and empty latent selection for image generation workflows.
Nodes to enhance & streamline prompts. Enhance using local LLM within ComfyUI, generate using random lines from text files, or randomly select text file prompts
Audio processing nodes for comfyui.
Stacked multi-LoRA loader for MiniMax H3 across bf16, fp8, int8 and w4a8 weights.
Workflow templates from ComfyUI Wiki - No custom nodes or dependencies included
Second monitor preview for video generations with interactive playback controls, color tools, generation tracking, and workflow-embedded snapshots.
custom node for ComfyUI to perform simple math operations
Simple ComfyUI node that predicts the score of an aesthetic image with SigLIP-based predictor.
ComfyUI memory cleanup nodes for VRAM and RAM management.
ComfyUI custom node that runs Codex CLI image generation and image editing workflows.
An extended Load Image node with an aspect-ratio-locked, WYSIWYG crop rectangle that outputs the cropped IMAGE and MASK.
Essential nodes for semantic image editing workflows with Kontext, Qwen Image Edit, and HiDream E1.1 models. Features automated iterative editing with Boyo Image Grab, organised dataset creation with paired image/prompt savers, and comprehensive file management tools. Perfect for batch processing, training data preparation, and streamlined semantic editing pipelines. Added Uncensored Qwen VL abliterated model node and downloader!
Video frame interpolation nodes for ComfyUI using BIM-VFI, EMA-VFI, SGM-VFI, GIMM-VFI, SPEED, and LDF-VFI.
FL ClearVoice - Audio enhancement nodes for ComfyUI. Features speech enhancement, super-resolution, denoising, and audio restoration using ClearVoice, Resemble-Enhance, and VoiceFixer models. Supports multiple backends with automatic model downloading.
Nodes: LoadImageFromBase64. Loads an image and its transparency mask from a base64-encoded data URI for easy API connection.
Unofficial implementation of [a/hallo](https://github.com/fudan-generative-vision/hallo) for ComfyUI
ComfyUI nodes that turn a panel layout (45+ presets, or hand-drawn) into per-panel generation dimensions, then composite the generated panel images back into one finished comic/manga page - including non-rectangular panel shapes.
Used to solve the OOM (Out Of Memory) issue caused by loading all frames of a video at once in ComfyUI. All nodes use streamingly, and no longer load all frames of the video into memory at once.
A ComfyUI custom node for integrating with Midjourney API.
Temporally apply visual LTX 2.3 LoRAs to selected regions of a single continuous video sampling run.
Provides tools for applying and removing fisheye lens effects from images.
ComfyUI wrapper node to test LaVi-Bridge using Diffusers
A robust custom depth estimation node for ComfyUI using Depth-Anything models. It integrates depth estimation with configurable post-processing options including blur, median filtering, contrast enhancement, and gamma correction.
ComfyUI custom nodes for ToriiGate-0.5, an image captioning model for anime-style and digital art.
Use power of three.js in 3d view on comfyui.
This is a ComfyUI custom node that helps generate images while preserving the text, logos, and details of e-commerce products.
Visual layout editor for MOSS-Image: draw boxes on a canvas and compile the official summary/items layout JSON.
It make any text2image create seamless patten
PBR texture generation diffusion model nodes. Generate depth maps and normal maps from baked textures using PBRFusion4.
ComfyUI-3D-Covert is a 3D File Format Conversion Extension based on 3dconvert.nsdt.cloud-API. It provides a node ConvertTo3DFormat currently available, is a powerful tool designed to streamline the conversion of 3D models between a wide array of file formats.