ComfyUI Custom Extensions
Browse 9,446 ComfyUI extensions and 73,215 custom nodes.
Repository contains CSV Styler - custom node for ComfyUI. It loads styles from styles.csv file (Automatic1111 Web-Ui styles.csv format) and combine them with provided positive and negative prompt. NOTE: 'styles.csv' should be placed in ComfyUI root directory (near main.py).
This plugin adds functionality to ComfyUI for detecting yellow tones in images, making it particularly useful for skin tone analysis and image color evaluation.
A custom ComfyUI node that integrates with the OpenAI Assistants API.
A high-performance image processing node pack for ComfyUI with 90+ nodes for filters, transforms, compositing, drawing, segmentation, and more.
Personal ComfyUI custom node toolkit — dynamic prompts, utilities, and more.
通过任意 OpenAI/Gemini 兼容网关调用 NanoBanana(Gemini) 文生图/图生图,支持最多14张参考图、批量并发生成与端点延迟测速。
A custom node for ComfyUI that enables direct posting of images, videos, and messages to Discord channels. This node seamlessly integrates your ComfyUI workflows with Discord communication, allowing you to automatically share your generated content.
Transcribe audio with a local Nemotron ASR (NeMo-Speech.cpp) server: batch, file, realtime WebSocket streaming, and a live microphone node, all with live inline text preview on the node.
Reusable system prompts as first-class ComfyUI assets: a persistent prompt library with Edit/New/Delete on the node, a selector that outputs the prompt as STRING, and a combiner that merges system+user prompts through configurable templates.
ComfyUI text nodes. Pre (preprocess comments, random, and de-emphasis), Show (show and persist text), and Debug (show and persist weights).
ComfyUI custom node: range-aware Concat/Combine/Average for CONDITIONING that correctly carries start_percent/end_percent across both inputs. Single unified node with mode dropdown, dynamic 1..N input slots, per-slot weights, and timestep-range intersection.
Comfyui custom node: Set of tools for pixel art palette.
Replace the voice in any audio with another voice (Seed-VC zero-shot voice conversion). Keeps music and background, exact length for video sync, one voice or two speakers each with their own voice. Standalone, models download automatically.
Custom ComfyUI nodes for integrating WaveSpeed AI API for image and video generation.
Nodes: Combine ZHGMasks, Cover ZHGMasks, ZHG FaceIndex, ZHG SaveImage, ZHG SmoothEdge, ZHG GetMaskArea, ...
Nodes: VAE Processor, VAE Loader, Background Remover
Ideogram 4 quad nodes for ComfyUI: layout, prompt builder, keyframe, per-panel inpaint, and timed keyframe-release.
Security scanner for ComfyUI workflows, custom nodes, and model files
A useful comfyui node named LoadArtistTag for selecting artist tags, including 1000+ single-artist tags and 300 mixed-artists tags.
Generates EU AI Act Article 50 disclosure records as PDF and JSON from the executed workflow. Three nodes: Creator, Broadcast, Project.
Clear current image selections, conventional previews and prompts without deleting history.
Combines texts line by line to a single text
A collection of custom nodes for ComfyUI.
Prompt history node for ComfyUI: appends prompts to a timestamped JSONL log and generates a Ditto-style HTML viewer (newest first, click to copy).
ComfyUI custom nodes for VoxCPM text-to-speech, voice cloning, multi-speaker generation, and LoRA training.
Nodes: Image Crop and Resize by Mask, Image Uncrop, Image Shadow, Optical Flow (Dip), Warp Image with Flow, Image Threshold (Channels), Finetune Variable, Finetune Analyze, Finetune Analyze Batch, ... Misc ComfyUI nodes by Rick Love
🧝 Interactive assistant that lives in your ComfyUI workspace. Reacts to queue events with animated sprites, voice lines, and text bubbles. Features multiple personalities: Diana (calm), Kitty (cute), Kerrigan (dark). Includes audio toggle, idle timer, and smart error detection.
Sigma scheduler for Qwen-Image 2.1 Turbo, Krea2 Turbo, Flux2 Klein, Z-Image Turbo. One slider, three tuned profiles.
ComfyUI node integrating Black Forest Labs FLUX.2 [pro] API via Floyo for text-to-image and image editing. (Description by CC)
RizzNodes for ComfyUI Welcome to RizzNodes, a collection of custom nodes for ComfyUI designed to streamline various workflows, from loading images and models in batches to dynamic prompt generation and memory management.
Load file paths from a folder by extension, then feed them into any ComfyUI node that accepts a path. (Description by CC)
Prompt Relay custom nodes for Wan2.2 I2V and LTX-Video 2.3 that connect multiple scenes in latent space to reduce character drift and improve consistency. (Description by CC)
Batch image-to-video without repeats: tracks which source images already produced a video, with a gallery, random pick, and rerun for any LoadImage workflow
shangxiaobao.com deployment platform, turn your comfyUI workflow into an APP on shangxiaobao.com
A custom node for ComfyUI that forces a connection between any input and output.
Helps filter and trim image batches for I2V chaining by ensuring consistent face identity and eye openness.
Process batches of images (e.g., video frames) with ultralytics bbox detection in ComfyUI. This custom node package extends the standard BboxDetectorCombined to handle batches efficiently, making it ideal for processing 300+ video frames.
Makes ComfyUI more sweet
High-quality text-to-speech ComfyUI custom node powered by Maya1 model
Automated checkpoint cycling node for ComfyUI that switches models on each execution with configurable modes (sequential, random, weighted sampling) and supports batch queuing. (Description by CC)
Create missing OGN-ModelManager checkpoint thumbnails from checkpoint-associated generated image folders.
Syncs checkpoint selection lists without restarting ComfyUI; updates backend Python type definitions and maintains compatibility with standard nodes and cycler nodes. (Description by CC)
A lightweight IMAGE preview tap node that sends an in-memory JPEG contact sheet without writing temporary preview files.
Pillow-based direct JPEG saver node for ComfyUI. Recommended image output node for HandpickerSuite.
2–4× faster generation by processing peripheral image regions at lower resolution while maintaining full quality in the foveal region.
Standalone Smart LM Loader, Smart Detection and bbox conversion, Registry Manager, verified model acquisition, and native or Docker LLM/VLM backends for ComfyUI.
Simple node to manage API keys and secrets for OpenAI, Anthropic, OpenRouter, Stability AI, etc... Allows you to keep your secrets out of your workflows. Zero Deps
ComfyUI Nodes for misc. use-cases
Implementation of xie-lab-ml/Zigzag-Diffusion-Sampling into ComfyUI, tested primarily with Anima video workflows.
This repository provides a set of custom nodes for ComfyUI, designed to improve workflow convenience when working with text, image, video, and audio data.
This repo make AI-Toolkit FlowMatch scheduler working on ComfyUI.
A ComfyUI custom node for bulk creative prompt generation that takes a seed idea and generates batches of distinct, shaped prompts for LoRA testing workflows. (Description by CC)
Imagen descripcion prompt system
Custom nodes for ComfyUI that integrate with the RaccoonAI local AI management system.
Load images with their generation data (PNG, JPEG, WebP; single file, by index from a folder or the newest) and read it back: ComfyUI prompt and workflow, A1111-style parameters (seed, steps, CFG, size, prompts, model, LoRA stack), LoRA hashes for Image Saver; Save image that collects the parameters from the graph by itself; automatic colour correction without neural networks (white balance, levels, vibrance, reference match, histogram). Based on ComfyUI-Crystools.
Custom node that takes an audio mp3 and an image, and creates a lip synced video from them.
A tool to reverse the frames of video type files (MP4, WebM, WebP, Animated GIF) using ComfyUI. It gets the image batch tensor and returns the same tensor, but flipped.
A ComfyUI custom node package for managing subprocesses within ComfyUI workflows.
Image labeler for ComyUI
Stamp a verifiable Mememage bar into an image and read it back — the image carries a key to a record you build from your own fields (your generation data, a creator name, a license, anything you choose).
AIO Pony Node.
Does the same thing as comfyui_image_metadata_extension but is slightly more modern, sending the output to Telegram and downloading unnecessary metadata at the same time.
Track model usage frequency to identify abandoned or underutilized models. Shows usage stats, file sizes, and highlights stale models in a sidebar panel.
This project provides custom safety checkers for image classification using Falcons AI and CompVis models. The safety checkers are designed to detect and filter out NSFW content from images.
ComfyUI node pack for DiffusionGemma — uniform-state discrete diffusion text generation, made instrumentable.
ComfyUI custom node that merges a batch of images into a single grid canvas.
Personal ComfyUI experiments and utilities repository. (Description by CC)
Dynamic multi-channel switch node with arbitrary input count, precise matching, default fallback, and lazy loading for ComfyUI. (Description by CC)
Custom-node extraction lane for HY-OmniWeaving support on top of stock ComfyUI. (Description by CC)
ComfyUI Node Implementation: TeaCache Acceleration Specifically Designed for the Lumina Model
Various quality of life -nodes for ComfyUI, mostly just visual stuff to improve usability.
Load workflow from civitai image.
Create a new group of nodes.
Load image and partially workflow with javascript.
Prevent sleep while running ComfyUI.
Move canvas to target on dragging connection.
Add command and comment in textarea. (e.g. // Disabled line)
Add keybindings to textarea.
image split comparison horizontally/vertically, real-time mouse interaction to adjust split ratios, bidirectional linkage between slider and mouse, and the ability to export comparison images with precise split lines.
ComfyUI nodes for LoRA training evaluation — XY grid sampling across checkpoints and prompts
ComfyUI video style transfer plugin based on TeleStyle for applying reference image styles to source videos. (Description by CC)
Extra nodes for ComfyUI, from the Winteraction workbench. Time works for you: WCheckpoint, WLoRA, WLoRA Stack, WSampler, WScheduler, WInt (seed), WFloat (cfg or denoise) and WSwitch have a control after generate and the carry that links them, so a queue of runs changes everything on its own, and WSave Image writes into every file name the values that made that image. One press of Run and the machine never stops. 34 nodes in all, for the classic canvas and for Nodes 2.0. With them WReso, WRatio, WFrame, WPrompt Rows, WRoute, WRoute Index, Wif, WIndex and WDifference; under WextraUI/H3 the 7 of the scene loop for MiniMax H3 Sync Sound, a film as a list of scenes with continuous audio; under Video, Audio and Foley WExtractor, WMount, WCutScene, WAudio Prompts, the Foley loop (LTX-2.3 video-to-audio, a prompt per scene, in one Run), WMixer to build the soundtrack in the timeline inside the node and WAonV, the sound on the picture. No ffmpeg to install.
Custom ComfyUI nodes to help with creating custom Yu-Gi-Oh! cards and modding the video games
ComfyUI custom nodes for RSA image encryption using asymmetric cryptography with key generation, encryption, and decryption functions via the cryptography library. (Description by CC)
A collection of utility nodes for ComfyUI that enhance workflow capabilities with advanced loaders and processing tools.
NODES:Animated Offset Padding, Animated Rotation & Zoom
Load numbered videos from a folder for batch workflows. Supports skip-missing mode for sequences with gaps.
A collection of useful custom nodes for ComfyUI workflows
One-click backup of the LoRAs that are active in your current ComfyUI workflow. Skips disabled Power Lora Loader rows, keeps folder structure, never overwrites unnecessarily.
Load an MP4/AVI video, scrub to any frame inside the node (drag & drop supported) and output the visible frame as an IMAGE or save it as JPG/PNG. Shows video and frame metadata.
ComfyUI node integration for the Camie tagger v2 model from Hugging Face for image tagging. (Description by CC)
High-quality multilingual TTS with voice cloning, emotion control, and support for Qwen3-TTS models (CustomVoice, VoiceDesign, Base).
Wrapper for 'resemble-ai/chatterbox'
ComfyUI custom node that selectively filters facial keypoints from DWPose/OpenPose output, allowing selective disabling of face parts (chin, eyebrows, eyes, nose, mouth outer/inner contours) for ControlNet workflows. (Description by CC)
A ComfyUI custom node for multimodal inference with Qwen3VL vision-language models accepting text, up to 4 images, and video in one node.
Meme Maker Node for ComfyUI.
HOTCUT nodes for ComfyUI — cloud image models (GPT Image 2, background removal) billed in rubles via the HOTCUT API. Sharp text, Cyrillic and faces that local models struggle with, right inside your graph.
NODES:Aesthetics, Aesthetics V2, Load AI Toolkit Latent Flux, Send_to_Eagle, Snek SQLite Prompt Logger
A custom node for ComfyUI that allows you to send images directly to [a/Eagle](https://jp.eagle.cool/).