ComfyUI Custom Extensions
Browse 9,420 ComfyUI extensions and 73,215 custom nodes.
ComfyUI plugin to monitor Dynamic VRAM usage
A model downloader plugin for ComfyUI
ComfyUI-denoise-mask-scheduler experimental approach involves selectively applying a denoise mask at each step during the inpainting inference process in diffusion models.
Photoshop custom nodes inside of ComfyUi, send and get data via Photoshop UXP plugin for cross platform support
Unattended multi-shot MiniMax H3 prompt chains for ComfyUI
This custom node allow controlling output without training. The reducing method is similar to [a/Spatial-Reduction Attention](https://paperswithcode.com/method/spatial-reduction-attention).
Drift Johnson's KokoroTTS nodes
ComfyUI-Neo-Nodes (Neo nodes,new version in Neo Studio): Neo Prompt Encoder and Neo Prompt Agent (LLM prompt enhance, generate, reverse, translate, classify), Neo Image Gen & Edit (skill-template text-to-image, editing and outpaint with Autogrow reference images), Neo Reference Grid, Neo Grid Split, Neo Bundle Expand, Neo H3 Video Director with Neo H3 Add Keyframe, Add Context and Add Guides, Neo H3 Segment Run, plus built-in image and video generation settings, Gallery asset library with LoRA tagging, Recipes one-click send, Model Hub downloads, workflow repair, and the Neo Studio and Backyard pages.
Neo Studio (ComfyUI kernel plugin): Neo Prompt Encoder and Neo Prompt Agent (LLM prompt enhance, generate, reverse, translate, classify), Neo Image Gen & Edit (skill-template text-to-image, editing and outpaint with Autogrow reference images), Neo Reference Grid, Neo Grid Split, Neo Bundle Expand, Neo H3 Video Director with Neo H3 Add Keyframe, Add Context and Add Guides, Neo H3 Segment Run, plus built-in image and video generation settings, Gallery asset library with LoRA tagging, Recipes one-click send, Model Hub downloads, workflow repair, and the Neo Studio and Backyard pages.
This custom node provides a smooth step function that normalizes LoRA values by enhancing elements above the mean while reducing those below it. Users can independently control both the LoRA strength and smooth step intensity to fine-tune their results, though the effectiveness may vary between different seeds and LoRA types.
A ComfyUI custom node enabling Flash Attention on legacy NVIDIA GPUs (Tesla V100, T4) that lack Compute Capability 8.0+ required by FlashAttention-2.
Neural Dreamer Nodes — ComfyUI nodes: variant selection with live preview and a hand-off to your own upscaler, a modular image loader, a character prompt builder, parameter bundles and more.
ComfyUI nodes for loading SDXL checkpoints with 16-channel latent format support. Includes basic loader, precision-controlled loader (fp32/fp16/bf16 for MODEL/CLIP/VAE), and dtype inspection utility.
This project adapts the dashscope([a/aliyun-bailian](https://bailian.console.aliyun.com)) api into ComfyUI.
125+ production-ready V3 custom nodes for ComfyUI covering file ops, S3 cloud storage, image/video/audio processing, model management (HuggingFace/Civitai/S3), OpenAI & Runpod integration, and workflow utilities.
Split PSD into processable layer images, rebuild with metadata, and merge multiple PSDs or images on a shared canvas for ComfyUI.
A plugin that enables one-click language switching between Chinese and English in ComfyUI without page refresh. (Description by CC)
This is a custom node for ComfyUI that allows users to perform Image-to-Image generation by calling the FLUX.1-Kontext-Dev model via the official ModelScope API.
This extension helps generate images through Illustrious.
Advanced mask editor for ComfyUI, non-destructive editing with professional selection tools, custom PNG brushes, embedded preview, and fullscreen editor.
This custom node pack for ComfyUI provides an advanced image processing workflow to achieve high-quality upscales with an extended dynamic range (HDR), mimicking the flexibility of a camera's RAW file. Note: this node will not correct AI image hallucinations, ideally it is used in the last pass from a good quality image (2000/3000px with an X2 or X4 model).
TextMakerPro is a ComfyUI custom node and browser-based text-layout editor designed to create stylized text images directly from ComfyU
A ComfyUI custom node that assembles natural-language character portrait prompts from structured fields (origin, age, build, skin, face, hair, wardrobe, camera & lens).
This plugin is mainly based on Minio, implementing the ability to read images from Minio, save images, facilitating expansion and connection across multiple machines.
A custom node designed for ComfyUI, allowing users to format the current date and time based on a specified format.
Generates PBR textures
The 'Eses Image Offset' node offers basic image offsetting capabilities within ComfyUI. It allows shifting image and mask content horizontally and/or vertically, with an option to wrap content around the canvas edges for a tiling effect.
ComfyUI nodes that render timelines on Reactor real-time video models.
ComfyUI-RH wrapper for full Bernini Diffusers inference
Add custom Metadata fields to your saved PNG files.
Triton sparse attention, MiniMax H3 memory optimizations, and experimental text modulation for ComfyUI. Continued by r-vage from Saganaki22's original integration; validated on Linux with RTX 4070 Ti SUPER (Ada Lovelace, SM89).
Play Doom in ComfyUI.
A collection of high-performance neural network-based Super Resolution models for ComfyUI.
A custom node for ComfyUI that allows users to add text overlays to images with customizable size, font, position, and shadow.
ComfyUI custom nodes for Miso TTS 8B, a Sesame-style CSM text-to-speech model with Mimi audio codes.
ComfyUI node for Seedream 4 image generation via Replicate API with support for text-to-image and image-to-image generation
A collection of custom nodes for ComfyUI featuring Gemini AI integration for video and image analysis, with a complete development environment using GPU-accelerated containers.
Toggle launch flags (like Sage on/off, fast accumulations, etc.) and restart ComfyUI from the UI. This can be especially usefull on RunPod or other cloud services but works locally too.
[BETA] A universal ComfyUI extension that enhances any input field with advanced prompt editing, auto-completion, custom chants, and save/load functionality.
Nodes:Image Square Adapter Node, Image Resize And Crop Node
An assorted reliquary of nodes: mismatched, stubborn, and deliberately indistinct; they do what is required; do not ask for particulars.
Custom nodes for ComfyUI by Eagleshadow.
This is a custom node to convert png images into color ASCII art. As noted below, multiple font sizes are used in the specification. The resolution of the generated file is set to be the same as the input image.
An experimental implementation of MagCache for SDXL
Save and load images with prompts and generation metadata. Batch image loading with metadata extraction for ComfyUI
A ComfyUI custom node that simulates Photoshop's 'Flatten Image' functionality.
SDR-to-HDR conversion with LumaFlux (arXiv 2604.02787): prompt-free, 8 steps, a frozen FLUX.1-dev plus 71M adapter parameters. Loader, converter with video stabilization, SDR preview, and save nodes for 16-bit PQ PNG and 10-bit HEVC HDR video. The paper repository doubles as the node pack.
Sculpt 3D meshes directly in the ComfyUI. Ten brushes, OBJ/FBX/GLB/glTF import, OBJ/GLB/STL export and a preview image.
Open ComfyUI subgraph nodes in a live popup canvas.
A translation node for users in Chinese Mainland. (Because of the network firewall in Chinese Mainland, many translation APIs cannot be used normally.)
LongCat_Avatar,an upgraded open-source framework for audio-driven human video generation.
ComfyUI is proud to present a new plugin designed to enhance user experience through seamless integration with Pillow, the powerful fork of Python Imaging Library (PIL). This plugin offers a suite of basic image manipulation tools that are easy to use and integrate directly into the ComfyUI framework.
Improves performance of large workflows by rewriting the submitted prompt graph before validation to constant-fold switch/selector nodes and optionally prune now-unreachable branches.
Adds a Navigator sidebar for keyboard-first workflow search and jump.
Nodes: Latent Mirror. Node to mirror a latent along the Y (vertical / left to right) or X (horizontal / top to bottom) axis.
A series of enhanced nodes related to ACEStep
Auto-rig a static mesh into a skinned .glb (skeleton + weights) - fully pure-Python glb I/O, no Blender/bpy subprocess, plus Mixamo/UE5 relabel, texture transfer, voxel post-process, and skin-weight cleanup.
NVIDIA DLSS 5 Neural Rendering and RTX Upscaling for ComfyUI.
A comprehensive ComfyUI node package that uses Microsoft's TRELLIS2 to transform video footage into 3D mesh reconstructions with advanced rendering effects.
Auto detect and process sensitive areas using YOLO with mosaic, blur, white out, and PSD export.
ComfyUI 智能显存管理节点
A ComfyUI node for lipsyncing videos using Sync.so API
volcengine comfyui api
Native ComfyUI BluePixel editing with split models and INT8 ConvRot conversion
Native ComfyUI V3 nodes for DreamX-Creator joint audio-video generation and causal 2x refinement.
Native Kandinsky 6 video/audio nodes and single-file INT8 ConvRot conversion
Standalone LLaDA-Image Base and Turbo AIO models with native ComfyUI model management, text/VQ generation and editing.
Klein4&9 MultiAngle Prompt is a ComfyUI node that lets you quickly set the camera angle and distance to the subject in your prompt without typing long descriptions manually.
Bridge for the standalone ComfyUI Model Downloader. Analyzes workflows, finds missing models via AI, and manages your archive to save SSD space.
Structured video direction and MiniMax H3 generation for ComfyUI Prompt Studio
A set of custom nodes that are using Anthropic's Claude models for describing images and transforming texts.
Some useful nodes for ComfyUI (model loader & extensions for EBU-LMStudio)
ComfyUI custom node for removing reflections from images using the WindowSeat model.
This custom node pack for ComfyUI provides a suite of advanced tools for performing 'invert-and-resample' workflows with Rectified Flow models like Qwen-VL. It allows you to take an existing image, convert it to a noisy latent using a guided inversion process, and then use that latent as a starting point to generate creative, high-quality variations.
A Figma-style layout and snapping engine for the ComfyUI canvas.
Japanese-assisted tag autocomplete for ComfyUI text areas. Supports Danbooru tags, bundled natural language tags, local CSV tag files, Japanese aliases, popup settings, and optional Danbooru CSV update from Hugging Face.
This extension simply connects the nodes and specifies the output path of the generated images to a manageable path.
Training-free activation steering for ComfyUI: read a transformer block's direction between two prompts and inject it with a single strength scalar. Open weights only.
Add a mobile-friendly web interface to ComfyUI.
ComfyUI startup optimizer and patcher
Use a local LM Studio server as the LLM/VLM backend in ComfyUI.
Winnougan and WINT8 custom nodes for ComfyUI — prompt encoder, CLIP loader, power LoRA loader, INT8 diffuser loader, resolution pickers, and more.
A professional video camera parameter estimation toolkit based on the VGGT model.
Comprehensive mobile UI enhancement for ComfyUI that transforms the desktop-focused interface into a touch-friendly experience optimized for iPhone, iPad, and Android devices.
This repository contains custom nodes for ComfyUI designed to help structure, filter, and generate text prompts using categorized tag definitions stored in a YAML file.
This plugin makes it easy to call RunningHub cloud workflows and AI apps from ComfyUI with a clean, intuitive interface
ComfyUI processes local real-time camera feed and provides real-time preview of the result.
About DeepSeek Chat API Go here to register and get the api-key [a/https://platform.deepseek.com/](https://platform.deepseek.com/) Then enter api_key in config.json
Official ComfyUI custom nodes and full-task workflows for ControlFoley, supporting controllable SFX generation, video-to-audio, text-to-audio, text-controlled video-to-audio, and audio-conditioned video-to-audio.
Make Cobra avialbe in ComfyUI. Cobra: Efficient Line Art COlorization with BRoAder References
Simple nodes for loading image files.Nodes that include a simple remote connection to Photoshop, a node that can overlay and preview an image with a mask, and a node that can load images directly from a file path.
A collection of utility nodes for ComfyUI, including audio/video processing, file uploads, and AI image generation.
Frontend-first workflow variant queue builder for ComfyUI
This pack provides two custom nodes for ComfyUI designed to find the center coordinates of masks and segments. They are useful for workflows where you need to precisely identify the location of objects or regions of interest.
Kandinsky 2/20B ComfyUI support with FP8/GGUF/blockswap support.
This is a custom node for ComfyUI that provides a dynamic 'Switch' for routing purposes. It allows you to define a list of named labels and select one, outputting the corresponding index and label name. This is useful for controlling the flow of your workflow based on a selection.
Professional audio editing nodes for ComfyUI — trim, measure, and inspect audio without leaving your workflow.
Unofficial implementation of KLing for ComfyUI
It's a super simple custom node for Comfy UI, to generate text, with a font size option. Useful for bigger labelling of nodes, helpful for wider screen captures or tutorials. Plus you can of course use the text within your generations.
NODES:TextNode, PreViewVideo, VideoSysNode.