ComfyUI Custom Extensions
Browse 9,420 ComfyUI extensions and 73,215 custom nodes.
A better sampler for Qwen Image 2.1: DPM++ 2M with adjustable denoised-history sharpening.
Meta SAM-Audio sound separation with text, time-span, and visual-mask prompts, ComfyUI model management, chunked processing, and optional Comfy Kitchen attention.
Mask and control-frame generation nodes for WanVideo/VACE workflows — supports 11 generation modes, source trimming, and merge-back with optical flow blending.
Takes the difference in between the positive and negative conditioning at the attention. NOTE: Will not work with Flux
Lightweight [a/PyAV](https://pypi.org/project/av/)-powered ComfyUI nodes to load/save multimedia.
editing activations in wanvideo
FiL Design ImageMind nodes for ComfyUI — image understanding and prompt generation (V3 API).
It is a simple HunyuanVideo block swap node for ComfyUI native nodes.
ComfyUI-native nodes to run First Order Motion Model for Image Animation and its non-diffusion-based successors. [a/https://github.com/AliaksandrSiarohin/first-order-model](https://github.com/AliaksandrSiarohin/first-order-model)
A collection of ComfyUI nodes. Including: WFEN, RealViFormer, ProPIH
Ripple Edit makes space in your workflow by moving nodes to the side as a group while keeping their relative position intact. Organize your workflow layout with ripple move and align compositions of nodes inspired by video editors. Frontend-only ComfyUI custom extension.
Basic Dynamic Prompts functionality with a Rich Text texbox for coloring.
ComfyUI extension to scan local Civitai models and check for newer versions.
Nodes:Image Gradient,Mask Gradient
Tag completion with a1111-sd-webui-tagcomplete style wildcard sub-selection, supporting CSV files and multiple wildcard sources. Features smart parsing, text overflow handling, and full compatibility with existing wildcard files. v2.0.0 with major wildcard workflow improvements.
This custom node for ComfyUI allows you to test combinations of samplers and schedulers. It generates a batch of generated images(latents), as well as RGB images, each one annotated with the specific combination used, performance timing, and several image quality metrics (Laplacian Variance, Gradient Mean, FFT-based sharpness) for our internal usage, but kept as they could may be useful for others to.
A ComfyUI custom node that automatically selects appropriate video resolution dimensions based on generation mode, aspect ratio, and quality settings. Designed to work seamlessly with video generation models and KJNodes image resize nodes.
A soft, Anima-friendly Perturbed Attention Guidance node for ComfyUI.
Fast int8, fp8 and NVFP4 diffusion models in ComfyUI on Apple Silicon: Metal int8 matmul, flash and sparse attention, and fused kernels for Krea 2, MiniMax H3 and LTX 2.5, plus fixes that let quantized models run on MPS.
A model cached-loader custom node for ComfyUI.
Excel spreadsheet-driven ComfyUI nodes that let you load models, values, and workflows based on saved rows in Excel. Great for organizing and switching between CLIPs, VAEs, LoRAs, and more.
A ComfyUI custom node that randomly selects a height and width pair from a list in a config file
Nodes:SaveImgAdv, CLIPTextEncodeWithStats. Collection of tools supporting txt2img generation in ComfyUI and other tasks.
Archive and organize your generated images into date-structured folders, and compress PNG images to JPEG/WebP with full workflow metadata preservation, saving 85-95% disk space.
The unofficial implementation of ZIM in ComfyUI
util nodes for custom node 'comfy-mecha'
Standalone ComfyUI node to grab the last N frames from a directory as a single batch with optional frame stride, designed for video continuation workflows and temporal consistency improvement.
ComfyUI custom nodes for VibeVoice ASR (Automatic Speech Recognition) with support for auto-labeling music/lyrics and high accuracy transcription.
Collection of custom nodes for ComfyUI designed to generate audio waveform visualizations ranging from real-time workspace previews to high-resolution images for video synthesis.
Diffusers custom_node
Custom_nodes for model in ComfyUI
Advanced multi-shot chaining for MiniMax-H3 audio+video with resumable cache, continuity controls, and unified model loading.
Batch samplers, schedulers, cfg, shift and steps tester custom node, automatic looping functionality for RES4LYF custom nodes
Load Anima Base and Anima 2.9B LoRAs on the inherited layers of Anima 3.8B.
Nodes:CLIP Text Encode (Batch), String Input, Batch String
NODES:HF ModelLoader, Show Images, Text2Image Inference, Decode Latent, Show CrossAttn Map, Show SelfAttn Map
ComfyUI nodes for RGB-X: Image decomposition and generation using intrinsic channels (albedo, normal, roughness, metallic, irradiance)
Adaptive sigma scheduling custom node for ComfyUI. Dynamically adjusts step sizes during sampling based on how the denoising direction changes.
This node is to fill image for outpainting(inpainting) Fill image using cv2 methods(cv2_ns, cv2_telea and edge_pad)
Nodes for loading and saving images with metadata in ComfyUI.
Pixstri is a custom plugin for ComfyUI designed to generate comic pages. It provides a hierarchical node system that allows you to create comic layouts with rows and frames, making it easy to design and preview comic pages within your ComfyUI workflows.
Generate human poses from text descriptions using T2P Transformer for ControlNet/T2I-Adapter workflows
Block Swap (RAM Offload) node for native ComfyUI MODEL - run DiT models larger than VRAM by streaming transformer blocks from system RAM.
Multi-area regional prompting nodes featuring a layered canvas mask GUI, greedy text encoder, and Gemini API persistent chat utilities.
A collection of ComfyUI custom nodes that allow to use most Diffusers pipelines and components in Comfy(Txt2Img, Img2Img, Inpainting, LoRAS, B-LoRAS, ControlNet...)
IndexTTS-2.5 的 ComfyUI 自定义节点
Makki's self custom nodes for ComfyUI.
The Face Cropper Node (MarwanFaceCropping) is a custom image processing node designed for ComfyUI. It takes an input image and crops it to a 2:3 aspect ratio, ensuring that most of the subject remains in the frame while maintaining the correct proportions.
Why make this node? Because I only need simple text related operations and don't want to install anything extra.
MTCLIPEncode: An extension for ComfyUI's CLIPTextEncode node, offering multilingual translation (using MarianMT) and prompt enhancement (using Ollama). Seamlessly translate your native language prompts into English and further optimize them for generating your desired images with Stable Diffusion. Supports Krita AI Diffusion.
Custom ComfyUI nodes used in Maxed Out workflows (SDXL, Flux, etc.)
Nuke-style image viewer for ComfyUI (beta): buffers, A/B compare, colour correction, video timeline, and EXR with full AOV support.
An enhanced extension for ComfyUI that records, visualizes, and extracts the actual execution path of a workflow.
Rust-powered ComfyUI node for correcting inconsistent AI-generated pixel art grids.
Nodes: InpaintMediapipe. This node provides a simple interface to inpaint.
Composite customizable speech bubbles and Japanese/multilingual text onto images in ComfyUI.
A specialized ComfyUI side-panel plugin designed for exploring 2D anime-style artist prompts with high-security defenses, automated style interaction, and prompt management.
[a/A3D](https://github.com/n0neye/A3D) is an AI x 3D hybrid tool that allows you to compose 3D scenes and render them with AI. This integration allows you to send the color & depth images to ComfyUI. You can use it as a pose controller, or scene composer for your ComfyUI workflows.
ComfyUI-NS-ManySliders is a custom node developed for ComfyUI that allows you to manipulate values using multiple sliders. With this node, you can easily adjust numerous numerical parameters intuitively, making it useful for various purposes.
A small set of preview and masking tools for ComfyUI
A ComfyUI extention for Kitten TTS
High-Performance Zero-OOM Multi-GPU Distribution & Memory Management Engine for ComfyUI — Layer Sharding, Device Splitting, and Dynamic VRAM Balancing for FLUX.2, LTX-Video, and Large DiT Models.
Segment Any Bedroom Interior is a Python-based project designed to segment furniture and objects within a bedroom image. The segmentation process uses RGB codes to accurately differentiate between various pieces of furniture, providing a precise mask output for each segmented object. This project is integrated with ComfyUI to allow easy and intuitive usage.
Several utility nodes for use with ComfyUI.
Consolidated Skeba reference, batching, bypass, and H3 Motion Context nodes for ComfyUI
A frontend extension for ComfyUI that allows you to open and edit subgraphs in a right-hand panel, instead of replacing the main canvas.
ComfyUI 实时性能监控插件,支持 CPU、内存、显存、磁盘和磁盘读写监控。A modern real-time CPU, RAM, VRAM, disk, and disk I/O monitor for ComfyUI.
MiniMax H3 for ComfyUI: one drag-and-drop input node covering t2va/fl2va/ref2va, plus step caching, partial decode and profiling.
This repository provides ComfyUI custom nodes to run WAN Video-As-Prompt (VAP).
Cleaner MiniMax-H3-Image text-to-image in ComfyUI: Steering and a Re-Noise Sampler against grainy fine texture, and a Resolution Selector up to 256 MP.
Idol by speridlabs natively in ComfyUI: one image to a 3D Gaussian splat on ComfyUI's own TripoSplat nodes, KSampler and memory management. Adds Idol's framing for photos, its predicted input camera and a precision node up to full fp32.
Find every cut in a video, fix the ones a frame off, pick the scenes you want and export them as clips: OmniShotCut detection, DINOv3 validation, a review viewer and frame-exact clip export, natively in ComfyUI.
A ComfyUI custom node for AI music generation using HeartMuLa models.
NVIDIA RTX Deblur & Denoise Nodes for ComfyUI\nThis extension provides GPU-accelerated deblur and denoise nodes powered by NVIDIA RTX technology via the nvvfx SDK.
The Polyhedron Suite for ComfyUI. Group-aware LoRA management: stack 10-25 LoRAs with per-group merge modes (SEQ/CONCAT/DARE), per-row decoupled CLIP strength, Trim & Resolve cleanup switches against multi-LoRA interference, thumbnails, trigger words, Civitai fetch, a dedicated Accelerator engine and a passive Merge Analyzer. Plus media I/O: a pinned-folder Media Loader (images, video, audio, batch modes) and a unified Save node for stills, video and audio. Plus sampling: a classic-wiring Sampler with a live preview decoder and a CLIP Text Encode that shares the Token Counter's count. Plus the Cine Timeline: a whole MiniMax H3 film as shots on a time axis, rendered clips as editable tracks.
Hunyuan 3D 2.1 image-to-3D generation nodes for ComfyUI with PBR texturing and model caching. Granular control for research.
A custom node that crops images to specified dimensions from a center point with optional offset
NODES:AIRedoon Qwen Model Loader, AIRedoon Translator, AIRedoon Image Caption, AIRedoon LoRA Stack, AIRedoon Image RGBA2RGB, AIRedoon Preview Text, AIRedoon Save Text, ... RedoonAi Tool Kit
A node to replicate [a/https://huggingface.co/spaces/latentexplorers/latentnavigation-flux](A node to replicate https://huggingface.co/spaces/latentexplorers/latentnavigation-flux)
Nodes for ComfyUI that extend the core functionality without adding extra dependencies.
A suite of advanced ComfyUI custom nodes for a complete 'Script-to-Screen' workflow, powered by local LLMs.
Nodes: Ruben-BilboX's PromptGeek Photo Prompt. This provides a convenient way to compose photorealistic prompts into ComfyUI. Post-Processing: adds various post processing effects. Bonus: Option to show a distant server shutdown menu.
A collection of custom nodes for ComfyUI that offers an extension to the existing Save and Preview Image/Video nodes allowing directly adding and previewing the files from your preferred cloud storage providers (S3/DropBox/Google Drive/...)
Custom node for ComfyUI. It adds additional metadata for saved images, ensuring compatibility with the Civitai website.
Audio batch creation, extraction, information, resample, mono and stereo conversion. Also cut, concatenate, blend (mix) and de/normalize. Join/split channels (stereo). Signal generator (`sine`, `square`, `sawtooth`, `triangle`, `sweep`, `noise`). Musical note to frequency. Audio downloader for quick workflows which downloads its example data.
Get metadata from image.
Load any image straight from your clipboard into a ComfyUI workflow with a dedicated node that handles clipboard image capture automatically.
The save node for ComfyUI reimagined. A session gallery on the node that survives restarts, hold-for-review triage that keeps your output folder clean, and one-click workflow recovery from any image - plus formats, metadata, credits, auto mask sidecar images (mediapipe), watermarks, a save-to-comfy-inputs button and more, all remembered between sessions.
GRAG-Image-Editing : Group-Relative Attention Guidance for Image Editing,you can try it in comfyUI
A tool for novice users in Chinese Mainland to call the huggingface hub and download the huggingface models.
Auto-tagging in sync with LoraTagsOnly, or use it as a simple text manipulation tool.Effortlessly flatten lists or output specific selections.Convert comma-separated text into individual toggle buttons for easy selective output.
Individual node snapshots and bounded link caching for smoother ComfyUI workflow navigation.
Trajectory-aware production execution optimizations for VDN-H3 / MiniMax H3 in ComfyUI
Auto-Rigging humanoid 3D meshes for Mixamo animations
A unified ComfyUI node for both LLM (text-only) and VLM (image-to-text) generation using GGUF models..
the official OpenBMB VoxCPM2 inference code as ComfyUI nodes
Original project: [a/link](https://github.com/tencent-ailab/PCDMs) Based on testing, the author's original images work very well, but using my own images generally requires some luck!
Custom nodes for building SAM3-centric editing pipelines inside ComfyUI, featuring mask merging, bounding box rendering, and DOM-based interactive editing. (Description by CC)
A ComfyUI custom node for the [a/Retro Diffusion API](https://retrodiffusion.ai/)