ComfyUI Custom Extensions
Browse 9,420 ComfyUI extensions and 73,215 custom nodes.
High-performance foveated sampler (FLSampler) for ComfyUI inspired by human vision; selective latent modification delivers cinematic clarity and micro-textures without generation overhead.
Get general description or specify questions to ask about images (medium, art style, background, etc.). Supports Chinese 🇨🇳 questions via MiniCPM model.
Generate OpenPose face/body reference poses in ComfyUI with ease. Made with 💚 by the CozyMantis squad.
If you want to draw two different characters together without blending their features, so you could try to check out this custom node.
ComfyUI reference implementation for [a/T-GATE](https://github.com/HaozheLiu-ST/T-GATE).
ComfyUI custom node for diff-based image merging
A ComfyUI extension that allows you to use some LLM templates provided by Ollama, such as Gemma, Llava (multimodal), Llama2, Llama3 or Mistral
This custom node for ComfyUI allows you to inject specific prompts at specific blocks of the Stable Diffusion UNet, providing fine-grained control over the generated image. It is based on the concept that the content/subject understanding of the model is primarily contained within the MID0 and MID1 blocks, as demonstrated in the B-Lora (Content Style implicit separation) paper. Features. Inject different prompts into specific UNet blocks Three different node variations for flexible workflow integration Customize the learning rate of specific blocks to focus on content, lighting, style, or other aspects Potential for developing a 'Mix of Experts' approach by swapping blocks on-the-fly based on prompt content
Advanced/Experimental VACE nodes for WAN video models in ComfyUI.
A tool for decomposing video storyboards, which can obtain storyboards and keyframes
This is a completely different set of nodes than Comfy's own KSampler series. This set of nodes is based on Diffusers, which makes it easier to import models, apply prompts with weights, inpaint, reference only, controlnet, etc.
a lightweight open-source node for ComfyUI, designed to simplify workflows while providing encryption protection for them.
A ComfyUI implementation of Latent Bridge Matching (LBM) for efficient image relighting. This node utilizes the LBM algorithm to perform single-step image-to-image translation specifically for relighting tasks.
Adaptive Prompts is a modern reimagining of dynamic prompts for ComfyUI. It lets you randomize, restructure, and clean up prompts with powerful wildcard and string tools. For the sake of consistency, I will still refer to them as Dynamic Prompts.
Removes the 2px pixel grid artifact from Qwen Image VAE decoded images with auto-calibration.
Spectrum-style hidden-state forecasting acceleration for native ComfyUI MOSS-Image.
Custom nodes for ComfyUI using BRIA's API.
Photoshop type functions and adjustment layers: 30 blend modes, Selective Color, Blend If, Color Balance, Solid Color Images, Black and White, Hue/Saturation, Levels, and RGB Splitting and Merging.
The ultimate A1111-style sidebar. Bring the classic UX you know and love into ComfyUI.
Comfyui custom node for [a/FunAudioLLM](https://funaudiollm.github.io/) include [a/CosyVoice](https://github.com/FunAudioLLM/CosyVoice) and [a/SenseVoice](https://github.com/FunAudioLLM/SenseVoice).
Temporal slit-scan / time-displacement effects for ComfyUI — angled, animated and dual-source slicing on a shared engine.
Canvas to use with ComfyUI
WildPromptor simplifies prompt creation, organization, and customization in ComfyUI, turning chaotic workflows into an efficient, intuitive process.
ComfyUI nodes for Sa2VA (Segment Anything 2 Video Assistant) - Advanced multimodal image and video understanding with segmentation.
Provide utilities for batch based video generation workflows (s.a. AnimateDiff and Stable Video Diffusion).
A few nodes to mix sigmas and a custom scheduler that uses phi, then one using eval() to be able to schedule with custom formulas.
Some useful custom nodes that are not included in ComfyUI core yet
Provides some features and nodes related to API calls.
Image captioning, prompt engineering, face processing and utility nodes for ComfyUI.
A video stabilization node for ComfyUI
A simple minimap in the bottom-right of the window showing the full workflow, left click to navigate
Nodes:Loader, Prompts, ImageTool, Wildcards, LoadEmbedding, ConditionText, SaveImages, ...
InpaintEasy is a set of optimized local repainting (Inpaint) nodes that provide a simpler and more powerful local repainting workflow. It makes local repainting work easier and more efficient with intelligent cropping and merging functions.
This is a TeaCache acceleration node for HunYuan Video, supporting the native node workflow for seamless upgrades. Simply choose the acceleration multiplier you want—currently, three levels are available.
The custom nodes of laksjdjf have been integrated into the node pack of cgem156🍌. NOTE:This includes the attention couple feature.
ComfyUI-Lightx2vWrapper is an inference wrapper for Lightx2v designed for use with ComfyUI.
PhotoDoodle: Learning Artistic Image Editing from Few-Shot Pairwise Data,you can use it in comfyUI
Unofficial ComfyUI implementation of [a/RAVE](https://rave-video.github.io/)
ComfyUI nodes based on the paper [a/FABRIC: Personalizing Diffusion Models with Iterative Feedback](https://arxiv.org/abs/2307.10159) (Feedback via Attention-Based Reference Image Conditioning)
Generate random prompts easily.
ComfyUI music production with YuE2 audio cover versions, SheetSage2 score-aware prompts, YuE2 and MiniMax Music 3 song generation, restoration, mastering, artwork and release metadata.
ComfyUI custom nodes for SoulX-Singer - Zero-Shot Singing Voice Synthesis
A custom node for ComfyUI that injects controlled noise into conditioning embeddings (like prompts) for a portion of the diffusion process.
A variety of custom nodes to enhance ComfyUI for a buttery smooth experience.
ComfyUI nodes to facilitate parameter/prompt keyframing using comfyui nodes for defining and manipulating parameter curves. Essentially provides a ComfyUI interface to the [a/keyframed](https://github.com/dmarx/keyframed) library.
ComfyUI-SoulX-Podcast is a custom node plugin for ComfyUI that packages the core functionality of SoulX-Podcast (long-form, multi-speaker, multi-dialect podcast voice generation) into a visual node workflow.
ComfyUI-CatvtonFluxWrapper provides ComfyUI nodes for diffusers implementation of Catvton-Flux.
Native photo deblurring nodes for ComfyUI, integrating HI-Diff and MISCFilter to reduce camera-shake and motion blur. Suitable for trying on handheld night photos, architecture and street scenes, signs and text, and portraits with mild to moderate camera shake. Supports original-resolution tiled processing, configurable overlap and blending, optional 8-way TTA, and official model downloads with SHA256 verification. Both models offer RealBlur-J, RealBlur-R and GoPro checkpoints, with Chinese and English instructions and example workflows. ComfyUI 原生照片去模糊节点,集成 HI-Diff 与 MISCFilter,用于改善手抖和运动造成的模糊。适合尝试夜景手持、建筑街景、招牌文字及轻中度抖动的人像照片。 支持原始分辨率分块处理、自定义重叠与融合、可选八路 TTA,以及官方权重下载和 SHA256 校验。两个模型均提供 RealBlur-J、RealBlur-R、GoPro 权重,附带中英文说明与示例工作流。
ComfyUI-EasyIllustrious is a custom node suite to make working with Illustrious models easier and more intuitive
NODES: ComfyUI-ImageFx, ComfyUI-Whisk, ComfyUI-Whisk-Prompts
MiniMax H3 FL2VA long-video continuation for ComfyUI with freeze-safe native masked video context, independent audio-tail preservation, flexible Qwen conditioning, and memory-bounded stitching
Node and workflow profiling. Find bottlenecks in your workflows. See trends over time.
SenseNova-U1: Unifying Multimodal Understanding and Generation with NEO-Unify Architecture
Adds a feature to ComfyUI that lets you nest nodes under a new node for better organization and simplification of repetitive patterns in workflows.
A ComfyUI custom node that can be added after any node to clean up VRAM and memory. Extracted from the original LayerStyle repository.
Nodes:LoadLoraWithTags. Save/Load trigger words for loras from a json and auto fetch them on civitai if they are missing.
MiniMax H3 generation for ComfyUI with an embedded 9-image, 3-video, and 3-audio reference gallery.
ComfyUI nodes for outpainting images with diffusers, based on [a/diffusers-image-outpaint](https://huggingface.co/spaces/fffiloni/diffusers-image-outpaint/tree/main) by fffiloni.
This is a Qwen-Image ComfyUI plugin implementation that can run the full version with 24GB VRAM
python and web UX improvements for ComfyUI: Lora/Embedding picker, web extension manager (enable/disable any extension without disabling python nodes), control any parameter with text prompts, image and video viewer, metadata viewer, token counter, comments in prompts, font control, and more! [w/'ImageFeed.js' from the custom scripts of pythongosssss is not compatible with this suite's ImageDrawer feature. Additionally, 'DynamicPrompts.js' and 'EditAttention.js' from the core, along with 'favicon.js' from the custom scripts of pythongosssss, are incompatible with advanced features of the suite. Please use the JNodes Extension Management setting in Settings > JNodes > Extension Management to disable these extensions by unchecking them to use the full functionality of the suite.]
Simple Qwen3-VL gguf LLM model loader
Custom loader module with optional 'any' parameter for flexible node connections and controlled model loading sequence to optimize VRAM management. (Description by CC)
Camera authoring, motion-scene editing, and model-native monitor pipeline for ComfyUI.
Dressup your models! This is my quick implementation of the SAL-VTON node for ComfyUI. Based on the paper [a/Keyu Y. Tingwei G. et al. (2023). Linking Garment with Person via Semantically Associated Landmakrs for Virtual Try-On](https://openaccess.thecvf.com/content/CVPR2023/papers/Yan_Linking_Garment_With_Person_via_Semantically_Associated_Landmarks_for_Virtual_CVPR_2023_paper.pdf)
Lightroom-style Camera Raw, Advanced AI Mask Editor (SAM 2.1/3, GroundingDINO, RMBG), and Crop-Pad-Outpaint (CPO) Editor for ComfyUI.
QR generation within ComfyUI. Contains nodes suitable for workflows from generating basic QR images to techniques with advanced QR masking.
ComfyUI nodes for Volcano Ark (Jimeng) API
ComfyUI node of [a/Kohaku's DanTagGen Demo](https://huggingface.co/KBlueLeaf/DanTagGen?not-for-all-audiences=true).
Input your desired output final resolution, it will automaticaly set the initial recommended SDXL ratio/size and its Upscale Factor to reach that output final resolution, also there's an option for 2x/4x reverse Upscale Factor. These all to avoid using bad/arbitary initial ratio/resolution.
Light-A-Video: Training-free Video Relighting via Progressive Light Fusion,you can use it in comfyUI
Official ComfyUI node for ACE-Step AI music generation via ACE-Step API
ComfyUI nodes for Black Forest Labs API Services
Custom node for ComfyUI. It adds additional metadata for saved images, ensuring compatibility with the Civitai website.
ComfyUI custom node for lightx2v
NKD Klein Presampling and Postsampling nodes for Flux Klein workflows in ComfyUI
ReCaVSR one-step streaming video super-resolution natively in ComfyUI: 4x on Wan2.2, block by block with its FlashDecoder. bf16 and int8 ConvRot, measured on an RTX 3090 and 4090, runs from 8 GB.
Advanced seed diversity enhancement for ComfyUI with intelligent noise injection and directional biasing.
Unofficial ComfyUI nodes for restart sampling based on the paper 'Restart Sampling for Improving Generative Processes' ([a/paper](https://arxiv.org/abs/2306.14878), [a/repo](https://github.com/Newbeeer/diffusion_restart_sampling))
This is a custom node of ComfyUI that categorizes tags outputted by tools like WD14Tagger, filters them by each category, and returns the filtered results.
Anima prompt workstation for ComfyUI with prompt presets, artist libraries, Danbooru and Gelbooru tag galleries, Civitai prompt import, and preview images.
Calling gemini2.0 at comfyui . The project will continue to organize good APIs!
This is a prompt to generate the word as the main function, but also reported the error node computing time dynamic display, the picture backward prompting, error assistant and translation and so on multifunctional nodes
A lightweight ComfyUI custom node pack for Qwen3-ASR, providing simple speech‑to‑text workflows with local model caching and optional timestamp output. Supports Qwen/Qwen3‑ASR‑1.7B and 0.6B, with HuggingFace/ModelScope download options and clean integration for ComfyUI pipelines.
Sigma-locked two-stage sampler and dual resolution selector for Krea 2 (and possibly other) workflows in ComfyUI.
Fast and accurate monocular depth estimation.
Custom ComfyUI nodes for Siray image/video models with dynamic schema-based inputs, authentication, and video streaming. (Description by CC)
Offline Chinese pose reference gallery for ComfyUI, with bundled assets and batch image nodes.
Content generation with open source models in comfyui via graq api implementation. [w/This repo is renamed from ComfyUI_GroqChat to ComfyUI_DW_CHAT. Please remove previous one and reinstall to this.]
A ComfyUI plugin based on SkinTokens (Newer Auto-Rig) from VAST-AI-Research
A nodepack designed to help you work with video from within ComfyUI that specializes in handling image batches efficiency and creating video loops.
Powerful video frame manipulation nodes for ComfyUI such as: efficient high quality batch scaling, arbitrary framerate resampling, seamless video loop tools, batch watermark composite, and more.
The Demofusion Custom Node is a wrapper that adapts the work and implementation of the [a/DemoFusion](https://ruoyidu.github.io/demofusion/demofusion.html) technique created and implemented by Ruoyi Du to the Comfyui environment.
Custom node for ComfyUI. It adds additional metadata for saved images, ensuring compatibility with the Civitai website.
A custom node for ComfyUI that allows you to interact with OpenRouter's chat/completion API, providing access to a wide range of models.
a collection of nodes to explore Vector and image manipulation
Simplified ComfyUI nodes for SCAIL-2 that automate video resolution adaptation, motion transfer/character replacement logic, colored mask generation, and segmented video output. (Description by CC)
This is an implementation of [MiniCPMv2_6-prompt-generator](https://huggingface.co/pzc163/MiniCPMv2_6-prompt-generator) by [ComfyUI](https://github.com/comfyanonymous/ComfyUI), including support for single-image caption, generate prompt by upload image and batch-images Prompt generation.
Anti-degradation nodes for MiniMax H3 in ComfyUI. Keep long videos and continuation chains sharp, with steady colour and no burn-in: Keystone and capstone hold clip 1's detail, Context Noise stops held frames from baking in, and Geiru Context / Anchoring handle clean joins. Includes Geiru Hybrid Windows for seamless long fast-motion renders and Geiru Window Director to build them window by window (approve, re-roll, rewind, blend). Example workflows included.
A visually stunning extension for ComfyUI that adds beautiful, customizable animations to both links and nodes in your workflow, with a focus on performance and customization. Includes an end-of-render animation and a text visibility tool for nodes. No extra packages are required, works with the latest version of ComfyUI, and should be compatible with most workflows. Larger workflows may experience performance issues, especially if you have a lot of nodes and are using a lower end system.
the comfyui custom node of [a/V-Express](https://github.com/tencent-ailab/V-Express) to make audio driven videos!