ComfyUI Custom Extensions

A Power-Lora-Loader-style dynamic LoRA stack node with a real on-disk folder browser, drag-to-reorder, optional CLIP, and per-lora split model/clip strengths.

ComfyUI startup optimizer and patcher

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.

Streaming text-to-video-and-audio (T2VA) nodes for MiniMax H3: a RAVEN model loader and a chunk-major RAVEN streaming sampler, built on ComfyUI's official H3 support.

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.

A ComfyUI custom node for formatting Anima prompts.

An intuitive, all-in-one node for ComfyUI that brings a powerful, layer-based regional prompting workflow directly into your graph. Say goodbye to managing countless Conditioning (Set Area) nodes and hello to drawing your creative vision.

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.

ComfyUI custom tool node collection with batch image loading, prompt queuing, text processing, video frame extraction, and prompt templating. (Description by CC)

NODES:TextNode, PreViewVideo, VideoSysNode.

An implementation of the DAViD tooling, a method for extracting depth, normals, and masks from an input image.

Builds a folder-based tree from prompt lines. Lets you organize prompts into folders. Also adds several configurable keybindings for text editing (line commenting, regional prompting areas). Splits the prompt into regions for Regional Prompting.

ComfyUI module for selecting and managing text prompts (poses) from a modular JSON library with visual preview. (Description by CC)

ComfyUI Interface for FudanCVL EffectErase

Nodes:Cascade Resolutions

Honest float32 out of any ComfyUI VAE. Decode without the [0,1] clamp and in float32 instead of bfloat16, measure what the stock path throws away, catch tiling artefacts, and write real float32 EXR. By Andromediastudio (andromediastudio.com). Ten nodes: Video QC (ingest report for generated video), VAE Decode (float32, no clamp), VAE Encode (float32), Image Range Stats, Image Compare (numeric), Tile Seam Check, Remap Range, Save EXR (float32), Load Audio (optional), Audio Latent Switch (generated / external).

Latent PNG utility nodes: save latents in your PNGs to reduce VAE baking

Unofficial Midjourney integration for ComfyUI

Professional compositing nodes for ComfyUI - bringing Nuke and Fusion workflows to AI-powered image generation

Apple Silicon fork of Continuity: one node for MiniMax H3, LTX 2.5 and the still families on Metal / MPS

A simple-as-possible custom node for ComfyUI to load LoRA models from a public URL.

Experimental utility nodes with a focus on manipulation of noised latents

A custom ComfyUI node that lets you upload generated images directly to your [a/BASE](https://getbase.app/) account — no manual downloads or re-uploads needed.

This is a node to remove background of human picture.

Compare ComfyUI parameter values without contact sheets — sweep numerics, samplers, or schedulers, then inspect each result full-res in MBQ Viewer.

A set of tools for generating and altering sigmas in ComfyUI.

Just a simple node to filter out NSFW outputs. This node utilizes [a/AdamCodd/vit-base-nsfw-detector](https://huggingface.co/AdamCodd/vit-base-nsfw-detector) to score the outputs. I chose this model because it's small, fast, and performed very well in my testing. Nudity tends to be scored in the 0.95+ range, but I've set the default to 0.8 as a safe baseline.

NODES:pose_extraction, Load_reference_unet, Load_denoising_unet, Load_Pose_Guider, Pose_Guider_Encode, DynamicPose_Sampler, load_pose_model, align

DEPRECATED - succeeded by Context-Anchored Tile Upscale (comfyui-contextanchoredtileupscale), a native rewrite of this pack's best path. Tiled SD upscaling with custom Guider support: use PerpNegGuider for negative prompts at CFG 1 and higher quality upscales. Context Only Overlap lets each tile use already-generated neighbor pixels as context without re-denoising them, greatly improving consistency and removing seams. Masked Upscaling re-diffuses only a masked region and skips untouched tiles, so parts of an image can be upscaled with different models or settings, or an inpainted patch refined to match. The anchor_context option holds unprocessed areas to the original image during sampling, preventing detail drift between sections and blending seams into the real image. Fork of ComfyUI_UltimateSDUpscale.

Transform text and images into immersive 360° 3D worlds using Tencent's HunyuanWorld-1.0 in ComfyUI with native FLUX architecture integration.

ComfyUI Gemma

ComfyUI Trajectory

Custom node for building keyframe timelines in Wan video generation with adjustable influence strength, supporting 1-8 keyframes. (Description by CC)

Works with any Depth Map and visualizes the applied version it inside ComfyUI.

Load LoRA files from any path on any mounted drive, with a tree-style file browser and LRU RAM cache.

DeepLearning is just a few clicks away!

A handy custom node for using Refiner (switching to a different checkpoint midway) When you specify the end of the base checkpoint, you can extract refiner_start which is end + 1. The output is fixed as an INT, so it can be passed to the handy custom node, Anything Everywhere? Since it only outputs a numerical value, it can also be used for other purposes.

Custom ComfyUI nodes for dynamic CFG control and realism enhancement, originally designed for Anima realism finetunes but works with any diffusion model.

Seamless inpainting for image-space models like Google Nano Banana and ByteDance Seedream 4. Fixes color mismatches using Poisson blending.

KSampler is provided, based on GPU random noise

Plays youtube videos in comfy. Use this node to share tutorials or renders. Youtube Playlists mode is also in Future Development in which you can add multiple youtube links and form a playlist which would be ideal for chained tutorials or lisitening and sharing songs playlists with others.

My Original Waifu is a tag-based prompt-transformation engine for ComfyUI designed for creators who want their original characters to feel consistent, expressive, and faithfully rendered across every scene. Instead of manually adjusting tags each time you generate an image, you define the essence of your waifu, her look, her outfits, the way she appears in different contexts, and the engine transforms your prompt to match those intentions. It stays out of the way and simply follows your rules with clarity and consistency. Your Waifu. Your Rules. Your Perfect Prompt.

A Tamagotchi that lives inside ComfyUI — feeds on your AI slop, evolves, feels love from API nodes.

A ComfyUI extension that collects generated images and opens them in a customizable local 3D gallery.

Comfyui-raw-image provides the ability to load raw image files for ComfyUI

Extra nodes for LTX-2 video generation in ComfyUI.

A connector between a Photoshop plugin and ComfyUI that enables the plugin to access ComfyUI's model list, LoRA trigger words, and preset workflows, and to directly send images and parameters from Photoshop to ComfyUI for image generation.

A ComfyUI extension to add spatial anchors/waypoints to better navigate large workflows.

A collection of forks, QoL nodes and utilities for ComfyUI

Context window sampling for LTXV / LTX-2 (LTXAV) video+audio models, including IC-LoRA guide splitting across windows. Self-contained; uses ComfyUI's context_handler and wrapper hooks (no core edits).

Small collection of typed primitive nodes.

Simple MCP Server for ComfyUI text to image Workflow - ComfyUI Node Integration , Based on lalanikarim/comfy-mcp-server work

Extremely inspired and forked from: [a/https://github.com/klimaleksus/stable-diffusion-webui-embedding-merge](https://github.com/klimaleksus/stable-diffusion-webui-embedding-merge)

Segment and Recognize Anything at Any Granularity.

ComfyUI custom nodes for mflux, including FLUX.2 workflows optimized for Apple Silicon.

Openpose editor implemented with konva.

Comprehensive media downloading and web scraping nodes for ComfyUI. Download from 1000+ websites including Instagram, Reddit, Twitter, YouTube, TikTok using gallery-dl, yt-dlp, and advanced web scraping. Features: Gallery-dl downloader (100+ sites), Yt-dlp downloader (1000+ platforms), Web Image Scraper with 15+ site-specific handlers, browser cookie authentication, MCP integration for Claude Desktop.

Decompose images into semantic RGBA layers using Qwen-Image-Layered model. Features AI-powered layer naming, PSD/TIFF export with proper layer structure, original resolution support, and VRAM management. Includes 18 nodes for layer decomposition, manipulation, and saving.

A custom ComfyUI node that draws parametric facial masks with black lines on white or transparent backgrounds. (Description by CC)

Video-first temporal foreground matting for ComfyUI using FudanCVL SAM2Matting with SAM2.1 and SAM3 variants.

Selector and Recourse. Presets & failsafes. Work flow. EXDYSA

These nodes are to edit the text vectors of CLIP models, so to customize how the prompts will be interpreted. You could see it as either customisation, 'one token prompt' up to some limitation and a way to mess with how the text will be interpreted. The edited CLIP can then be saved, or as well the edited tokens themselves. The shared example weights does not contain any image-knowledge but the text vector of the words affected.

A tiny patch altering Krea2 Turbo's attention mechanism based on the Krea-2-Turbo-Projector-Scale-LoRA for improved balance between creativity and prompt following.

This custom node set for ComfyUI provides a DatasetBatchNode for automated, sequential processing of datasets, particularly useful for iterative training or batched image/video generation workflows.

A ComfyUI plugin to simplify loading and managing custom nodes with a sidebar interface.

MiniMax Music 3 inference, MOSS dataset preprocessing, and LoRA training for ComfyUI

A powerful node for browsing and importing media from Pixabay directly within ComfyUI.

A ComfyUI node for uploading generated content to YouTube. [a/Buy me a coffee](https://buymeacoffee.com/chao.k)

qwen3vl API Supports batch image or video reverse inference.

Nodes:Incrementer, Width and height from aspect ratio, Width and height for scaling image to ideal resolutio. A simple set of tooling nodes.

This is a custom node for ComfyUI that allows you to download and apply LoRA models directly from Civitai/HuggingFace. It simplifies your workflow by managing LoRA downloads automatically, so you don't have to manually download files and place them in the correct folder.

Tiling and text utility nodes for VFX production pipelines in ComfyUI.

Long-video context windows, sequential sampling, and seam correction for ComfyUI Wan Animate2.

A1111-style PNG Info for ComfyUI: load an image, read its generation metadata as human-readable text (A1111 parameters and ComfyUI embedded workflows), and reuse positive/negative/seed/steps/cfg as typed outputs.

Audio Mixing node for ComfyUI

MUST HAVE feature for Cloud Installs. Single-Click Install Button to install Missing Model Files for Workflows. Supports 'extra_model_paths' via dropdown selection. Easiest way to use ComfyUI Ever! Instead of clicking the 'Download' button, click the 'Install' button instead! This easy 'Install' button securely has the server download the missing file(s) and install the file into the correct directory on the server. Why? You ever wanted to try out a workflow but didn't want to spend the 15 minutes to download the files and put the files where they need to be? Wait no more, use 'Model Installer' and click the 'Install' button. Enjoy.

Lightweight OpenAI-compatible LLM nodes for ComfyUI, with LM Studio, Ollama, Responses API, image input, MCP tools, and local MCP runners.

A ComfyUI node that extends an image vertically to add a label either above or below it.

Insert image to specific location of another image

Video editing nodes for ComfyUI with lazy frame-by-frame decoding for large video support.

Simple node to handle scaling of YOLOv8 segmentation masks

ComfyUI nodes for stitching a still frame onto the start or end of an IMAGE video batch and trimming extra tail frames with optional audio sync alignment.

Turn any image into clean, grid-aligned pixel art inside ComfyUI, using the compiled SpriteFusion Pixel Snapper engine — plus color controls, classic video-game palettes, a pixel-perfect resizer, and an interactive preview with a magnifier.

Nodes:Post to Discord w/ Webhook

Ultra fast video frame interpolation for ComfyUI using RIFE with TensorRT acceleration. Auto-detects CUDA version, installs matching TensorRT wheels, and downloads RIFE models automatically.

A simple wrapper for Fooocus's inpainting code, designed to replicate its outstanding results. Future improvements for a more refined and lightweight version are planned.

ComfyUI node that applies effects to generated images like JPG compression, pixelation, color grading, film grain, vignette, blur, and optional lighting matching with a reference image, and many more