ComfyUI Custom Extensions

Try Off for ComfyUI using Flux and CatVTON.

Using Spark-TTS in Comfyui. Spark-TTS: An Efficient LLM-Based Text-to-Speech Model with Single-Stream Decoupled Speech Tokenss

Nodes:Fit Size From Int/Image/Resize, Load Image And Resize To Fit, Pick Image From Batch/List, Crop Image Into Even Pieces, Image Region To Mask... A simple set of nodes for making an image fit within a bounding box

ComfyUI node for [a/Open-Sora-Plan](https://github.com/PKU-YuanGroup/Open-Sora-Plan)

ComfyUI custom nodes for the Civitai Orchestration API — run Civitai recipes (image/video/audio generation, training, analysis) from inside ComfyUI

ComfyUI node for Gemini Web API - text-to-image, image-to-image, and chat with multi-image input support

Simple set of nodes to assist with prompt generation for NOOBAI XL / Illustrious models

Parler-TTS is a lightweight text-to-speech (TTS) model that can generate high-quality, natural sounding speech in the style of a given speaker (gender, pitch, speaking style, etc)

An implementation of NSFW Detection for ComfyUI

An independent memory management custom node for ComfyUI. Provides Distorch memory management functionality for efficient GPU/CPU memory handling.

Load your model with image previews, or directly download and import Civitai models via URL. This custom ComfyUI node supports Checkpoint, LoRA, and LoRA Stack models, offering features like bypass options.

Based on the original repository [a/https://github.com/spacepxl/ComfyUI-Florence-2](https://github.com/spacepxl/ComfyUI-Florence-2), the model loading and storage methods have been improved, and sd3 has been newly added with enhanced speed and accuracy.

A ComfyUI custom node based on ByteDance MegaTTS3 MegaTTS3, enabling high-quality text-to-speech synthesis with voice cloning capabilities for both Chinese and English.

Extended model list for NAG integration in ComfyUI

A janky implementation of Sonar sampling (momentum-based sampling) for ComfyUI.

ComfyUI node for Lissajous camera shake effect that simulates natural handheld camera movements to enhance digital human video realism. (Description by CC)

Call LLM and VLM in a simple way using the OpenAI API standard from ComfyUI

Nodes:Load ICLight Model,Apply ICLight,Simple Light Source,Calculate Normal Map

Process text and images with GGUF LLMs in ComfyUI using llama.cpp, including Qwen3-VL, Qwen3.5, Qwen3.6, Gemma 4, and gpt-oss.

Run Qwen3-VL, Qwen3.5, and Qwen3.6 GGUF models in ComfyUI with llama.cpp.

This custom node for ComfyUI provides watermark detection capabilities using a YOLO model trained by [a/fancyfeast](https://huggingface.co/fancyfeast), the creator of JoyCaption. The model is originally hosted at [a/Hugging Face Space](https://huggingface.co/spaces/fancyfeast/joycaption-watermark-detection).

A ComfyUI custom node designed to eliminate pixel drifting, stretching, and minor cropping inconsistencies between a source image and an edited/upscaled version.

This repository provides utility nodes for defining inputs and outputs in ComfyUI workflows. These nodes are essential for running ShellAgent apps with ComfyUI, but they can also be used independently to specify input/output variables and their requirements explicitly.

This custom node integrates the [a/Step1X-Edit](https://github.com/stepfun-ai/Step1X-Edit) image editing model into ComfyUI. Step1X-Edit is a state-of-the-art image editing model that processes a reference image and user's editing instruction to generate a new image.

A HiDiffusion node for ComfyUI.

A pose remapping node with support for joint locking, motion-aware scaling, and optional easing - perfect for animation refinement and mech rig control.

Unofficial implementation of [a/PixArt-alpha-Diffusers](https://github.com/PixArt-alpha/PixArt-alpha) for ComfyUI

An image gallery node that displays previews for images in the ComfyUI/input folder for easy selection.

This is an ComfyUI implementation of LightGlue to generate motion brush

Node to enable WhatsApp in ComfyUI.

Allows to sample without generating any uncond with Stable Diffusion!

The OpenAI FM TTS node is a custom node for ComfyUI that seamlessly integrates the OpenAI FM Text-to-Speech service into your audio workflows. This node allows you to easily convert text to speech with a variety of voices and emotional styles directly within ComfyUI.

High-fidelity, zero-OOM tiled upscaler with adaptive processing for Flux.

GeekyRemB is a powerful suite of image processing nodes for ComfyUI, offering advanced background removal, animation, lighting effects, and keyframe-based positioning. Built on the rembg library with additional capabilities for chroma keying, mask refinement, realistic lighting, shadow generation, and dynamic animations.

Custom nodes with a touch of extra UX, including: history for primitives, JSON manipulation, logic switches with visual feedback, LLM chat... and more!

Warning, uses experimental package `comfy-env` to attempt a one click isolated install. Will download and use pixi package manager. Hunyuan3D-Part nodes for ComfyUI: 3D part segmentation with P3-SAM and high-fidelity part generation with X-Part diffusion.

The 'Eses Image Resize' node offers comprehensive image resizing capabilities within ComfyUI. It supports various scaling modes including scaling by a specific ratio, target megapixels, or directly to fixed dimensions. The node provides framing options to handle aspect ratio changes, allowing users to 'Crop to Fit' (fill) the target frame or 'Fit to Frame' (letterbox) the image with a customizable fill color. It also generates and outputs a corresponding mask, with control over the letterbox area's color (black or white) within the mask.

Basic Python functions for manipulating data that every programmer is used to. Currently supported: BOOLEAN, FLOAT, INT, STRING and data lists.

A ComfyUI node for pixel art scaling. Automatically detects the pixel scale using an edge-aware method (Sobel filter + voting on tiles) and downscales the image to that pixel size, reducing color palette.

Convert 2D images and videos to 3D SBS (side-by-side) format

Optional NSFW content expansion for ComfyUI-Photoreal-Prompt-Builder with explicit pose, action, and group action expansions.

ComfyUI extension that adds undo (and redo) functionality.

让 ComfyUI 像 WebUI 一样写提示词：中文翻译、Tag 自动补全、LoRA 管理、区域提示词控制，支持 Anima 与 SDXL 工作流。

Extensions: subgraph, setget, multiReroute

Apply screentone to line drawings or colored illustrations with diffusion models.

A ComfyUI implementation of Facebook Meta's [a/AITemplate](https://github.com/facebookincubator/AITemplate) repo for faster inference using cpp/cuda. This new repo is behind the old version but is a much more stable foundation to keep AIT online. Please be patient as the repo will eventually include the same features as before. NOTE: You can find the old AIT extension in the legacy channel.

Custom nodes for integrating Flux models with the BFL API.

A suite of experimental custom nodes that allows access to Google's 1P models like Veo, Imagen, Nano Banana, Gemini, Virtual-try-on, Lyria

a ComfyUI plugin for previewing latents without vae decoding. Useful for showing intermediate results and can be used a faster 'preview image' if you don't wan't to use vae decode.

Play with your pet while your workflow generates!

A theme pack for ComfyUI that transforms the nodes into the best 2000's aesthetics styles while preserving full functionality.

OpenVINO node is designed for optimizing the performance of model inference in ComfyUI by leveraging Intel OpenVINO toolkits. It can support running model on Intel CPU, GPU and NPU device.

This is a collection of custom nodes for ComfyUI that I made for some QOL. I will be adding much more advanced ones in the future once I get more familiar with the API.

Grounding for dummies, simplest workflow

ComfyUI custom node for CorridorKey-style edge-aware coarse mask refinement.

WarpFusion workflow wrapper for ComfyUI

Leverages Qwen 2.5/3 VL for prompt inversion & caption generation.

ComfyUI-Hunyuan3D-2.1 is now available in ComfyUI, Hunyuan3D-2.1 is a scalable 3D asset creation system that advances state-of-the-art 3D generation through two pivotal innovations: Fully Open-Source Framework and Physically-Based Rendering (PBR) Texture Synthesis.

ComfyUI integration for Pollinations API - Generate images and text based on user prompts

DeepExtract is a powerful and efficient tool designed to separate vocals and sounds from audio files, providing an enhanced experience for musicians, producers, and audio engineers. With DeepExtract, you can quickly and effectively isolate vocals or instruments from mixed audio tracks, facilitating tasks like remixing, karaoke preparation, or audio analysis.

Nodes:TextToSpeech, phy_3_conditioning, SaveAudioToDaVinci, SaveImageToDaVinci. NOTE:In order to use DaVinci node you must have DaVinci Resolve Studio connected to the API. For more information check the help seciton in DaVinci Resolve Studio HELP>DOCUMENTATION>DEVELOPER. It will open a folder, search for scripting and the for README.txt file, the API documentation.

This is an extension node for ComfyUI that allows you to load frames from a video in bulk and perform masking and sketching on each frame through a GUI.

All-in-one ComfyUI node that automatically generates arbitrarily long SCAIL-2 videos by internally looping chunked workflows with frame anchoring and continuous stitching.

Data research, preparation, and manipulation nodes for model trainers and artists.

A super simple node that outputs official zimage resolutions as a latent

Super LoRA Loader and ND Super Selector overlays for ComfyUI.

A custom node that adds a UI element to the sidebar allowing easy access, navigation, and use of a massive collection (100+) of LECO (Slider) LoRAs. LECOs are an amazing tool to generate variance in your output with a minimal impact to consistency, i.e. deviating form your prompt. They can also allow you access to control parts of your image without taking up CLIP space, saving your token weights for more valuable keywords. If you haven't used them, there's never been a better time to try!

An interactive image adjustment node for ComfyUI, with an easy-to-use graphical interface and realtime preview.

GeminiBanana is a custom node for ComfyUI based on the Gemini API. It allows you to call Gemini inside ComfyUI workflows to generate text, parse images, or perform multimodal interactions, greatly enhancing workflow automation and creative capabilities.

Advanced tiling plugin for ComfyUI solving VRAM limitations in 4K/8K+ image generation using intelligent tiling and seamless weighted fusion with multiple blending modes.

Interactive A/B image comparison node with a draggable slider to reveal one image over another. Includes difference and other blend modes for more detailed analysis, allowing one to spot changes in similar images. Node also outputs a passthrough image of input A, and a grayscale difference mask.

This node wraps the flux fill model as ComfyUI nodes. Use NF4 flux fill model, support for inpainting and outpainting image. Compared to the flux fill dev model, these nodes can use the flux fill model to perform inpainting and outpainting work under lower VRM conditions.

ComfyUI-AutoSplitGridImage is a custom node for ComfyUI that provides intelligent image splitting functionality. It combines edge detection for column splits and uniform division for row splits, offering a balanced approach to grid-based image segmentation.

Tired of forgetting and misspelling often weird names of embeddings you use? Or perhaps you use only one, cause you forgot you have tens of them installed?

A collection of image processing extension nodes for ComfyUI.

Generating seamless 360 degree panoramic image through text or perspective image.

Tile Splitter and Stitcher nodes for SeedVR2 upscaling in ComfyUI

A series of functional custom plugins, including Blender Bridge, Extract the value of any list type data, read text files, etc.

Five-node DiffusionGemma Director Assistant bridge for ComfyUI LTX workflows.

Nodes: SDXLRecommendedImageSize, JjkText, JjkShowText, JjkConcat. A set of custom nodes for ComfyUI - focused on text and parameter utility

comfyui-OreX is a set of nodes for using LLM models via a free API.

Data manipulation and retrieval nodes. Custom CLIP Text Encode with additional input and output convenience ports. Sum Wrap, a simple integer roll-over iterator. Custom LoRA Loader to extract and view meta-data and trigger phrases. Green Box is a node that can be chained to inject prompt fragments, iterating through a list. Python is a node that allows complete control, manipulation and creation of data, write your Python code inside this node. Multiple input/output ports referenced by name.

This custom nodes enables Stable Projectorz to work with ComfyUI Directly.

A communication bridge between Photoshop plugins and ComfyUI servers, enabling AI image generation by connecting PS plugins to Comfy servers for model inference and image rendering.

A ComfyUI implementation of SegviGen, providing precise 3D part segmentation using SegviGen.

A custom node for ComfyUI that integrates Ovi for synchronized video+audio generation from text or image inputs.

ComfyUI custom node: local llama.cpp → structured Ideogram 4 JSON prompt with thinking/reasoning support

This repository hosts a Text-to-Speech (TTS) application that leverages Whisper Speech for voice synthesis, allowing users to train a voice model on-the-fly. It is built on ComfyUI and supports rapid training and inference processes.

This ComfyUI node lets you browse the Civitai gallery directly within the interface, featuring infinite scroll, advanced filters (including NSFW), and favorites management. It also allows you to retrieve prompts, metadata, and images/videos to seamlessly reuse them in your workflows.

A set of nodes to use Reference UNets

ComfyUI node that reserves GPU VRAM before workflow execution to prevent OOM crashes with manual or auto mode options.

Local AuK Base and AuK-Flash speech generation, voice cloning, editing, enhancement and separation. Uses ComfyUI model management, attention and quantized operations.

Write and preview GLSL and p5.js shaders directly inside ComfyUI with a full code editor.

Nodes to perform tensor operations in ComfyUI

ComfyUI banana Nodes: Seamlessly integrate banana models into ComfyUI for enhanced image generation and image edit capabilities. Features include image generation and banana image edit.

Enhanced the comfyui savevideo node to support previewing and saving videos containing alpha channels.

A beautiful theme extension for ComfyUI that adds festive touches with dynamic backgrounds, snowfall effects, and animated node connections

a comfyui custom node for [a/GPT-SoVITS](https://github.com/RVC-Boss/GPT-SoVITS)