ComfyUI Custom Extensions
Browse 9,420 ComfyUI extensions and 73,215 custom nodes.
An extension for ComyUI that enables saving and loading node presets using the node's context menu. Right click a node and choose 'Presets' from its context menu to access the node's presets.
A custom LoRA loader node for ComfyUI with advanced block-weighting controls for both SDXL and FLUX models. Features presets for common use-cases like 'Character' and 'Style', and a 'Custom' mode for fine-grained control over individual model blocks.
A simple node to round an input image up (pad) or down (crop) to the nearest integer multiple. Padding offset from left/bottom and the padding value are adjustable.
A custom node pack for ComfyUI that aligns and retargets skeletal pose data from driving videos to match reference images, enabling consistent motion transfer. (Description by CC)
Nodes:SimDATrain, SimDALoader, SimDARun, VHS_FILENAMES_STRING_SimDA
DiffHDR (LDR to HDR reconstruction for images, videos and HDRI panoramas) on ComfyUI's native Wan2.1-VACE.
A modern, powerful, and easy-to-use Prompt Manager extension for ComfyUI.
A simple ComfyUI custom node extension that provides utility nodes for converting between integer and float values.
Modular procedural prompt generation — wildcards, fixed values, combines, derivations and constraints resolved in a context pipeline, with a manager UI for a reusable module library.
Advanced metadata management and multi-format image save system for ComfyUI. Save/load with metadata for PNG, JPEG, TIFF, PSD (with layers), SVG, and WebP. Features XMP/TXT/JSON sidecar files, embedded EXIF/XMP, MWG-RS regions for face metadata, workflow tracking, LoRA/embedding detection, runtime execution hooks, and Automatic1111 compatibility. Preserve layers, alpha channels, and complete provenance across all formats.
A modular character prompt generator for ComfyUI
This is a custom node for ComfyUI that embeds a simple web browser directly into the interface. It allows you to browse websites, find inspiration, and load images directly, which can help streamline your workflow.
Measure and repair face drift after an image edit: RAFT optical flow inside a Grounded SAM 2 face mask, and a smooth warp that moves the features back.
implementation easyanimate with llama3-8b-6bit instruction LLM generation prompt help
A specialized toolkit for AI-based image captioning (reverse prompting), prompt enrichment, and automated batch generation. It streamlines LoRA dataset organization with native GGUF and Ollama support.
Custom nodes: FSLGeminiChat, FSLGeminiGenerateImage, Transparent Background helpers, and more.
From F5-TTS for Thai language only เอาไว้สังเคราะห์เสียงภาษาไทย ตัวอย่างการใช้ https://youtu.be/Hpa2dnaRCmA
ComfyUI-GTSuya-Nodes is a ComfyUI extension designed to add several wildcards supports into ComfyUI. Wildcards allow you to use __name__ syntax in your prompt to get a random line from a file named name.txt in a wildcards directory.
Turns a short idea into a checked Danbooru-tag prompt for the Anima image model: a local LLM writes the tags, a builder validates them against a dictionary and assembles the prompt.
Simple camera prompt builder node for Qwen-Edit-2509 Multiple-Angles in ComfyUI.
The ComfyUI-FilePathCreator is a custom node extension for ComfyUI designed to generate dynamic filenames based on user-defined parameters. This node helps streamline the process of creating organized and timestamped filenames, ideal for saving output files in a structured manner.
Experimental Normalized Attention Guidance node for Anima Preview 3 in ComfyUI.
A simple, lightweight AMD GPU monitoring tool for ComfyUI that displays real-time information about your AMD GPU directly in the UI.
ComfyUI Blender Wrapper lets you call Blender in headless mode from ComfyUI (or any Python environment), run reusable pipeline stages stored inside files, and pass configs in/out via JSON. It's meant for tech-art / asset-pipeline tasks like cleanup, decimation, unwrap, baking, rigging, etc.
This is an implementation of [a/Qwen2-Audio-7B-Instruct-Int4](https://github.com/QwenLM/Qwen2-Audio) by [a/ComfyUI](https://github.com/comfyanonymous/ComfyUI), including support for text-based queries and audio queries to generate captions or responses.
A simple custom node to use reflect padding mode in the conv layers of VAEs.
Direct Higgsfield API integration for ComfyUI — multi-model image and video generation with Soul 2.0, Seedream 4, Kling 2.1, Seedance, and more. Use your own API key.
The most powerful API workflow collection available, covering a wide range of models.
Video and batch text workflow nodes for ComfyUI.
Node to automate batch generation with randomize prompts from text files. It mimics Forge's functionality, allowing you to combine text elements and LoRA. The node supports writing LoRA in any order within a text file using formats like <lora:name:1.0> or <lora:name:unet=1.0:te=0.75>, without needing separate nodes. The node understands LoRA names in Forge's style, when the name is not the filename, but the internal name from the metadata.
Publish a ComfyUI generation to Flow Production Tracking as a Version, with the model, prompt, seed and workflow that made it, and load a Version's media back into a graph.
Unofficial wrapper for Hunyuan PromptEnhancer
Aesthetic-focused custom nodes for ComfyUI with sophisticated style and prompt management
Aesthetic-focused custom nodes for ComfyUI with sophisticated style and prompt management
This extension is a node that directly expands the functionality of KSampler, rather than being in the form of a custom node. [w/Workflows created using this feature are not compatible with other users.]
StableDelight: Revealing Hidden Textures by Removing Specular Reflections
Essential Nodes Pack for Images Processing for ComfyUI, with a live embedded preview inside the node (no queue) for supported chains.
Nodes for ACE-Step 1.5 and XL: GGUF loaders, a multi LoRA/DoRA/LoKr loader with per-layer scaling, an experimental chord injector and a multi-audio comparison utility. Work in progress.
NDDG Great Nodes - ComfyUI custom nodes featuring Great Conditioning Modifier with advanced semantic drift, token dropout, gradient amplification and other conceptual manipulation tools. (Description by CC)
This is a ComfyUI integration of the [a/Dia TTS model](https://github.com/nari-labs/dia/). Many thanks to nari-labs for their fantastic work.
Smart dynamic layer swapping between GPU and CPU for optimal inference performance with comprehensive mixed precision handling and copy-compute overlap optimization. Enables running much larger models on limited VRAM setups.
WhisperCPP for ComfyUI - whisper.cpp with full CPP params, alignment, diarization
A custom node for ComfyUI that provides advanced transcription, alignment, and diarization capabilities using the WhisperX library.
A visual, interactive Lift, Gamma, Gain color correction node for ComfyUI, tailored for precise color grading workflows.
Drop-in workaround for LiteGraph-mode widget width bug in ComfyUI frontend. (Description by CC)
UI Decorators is a set of 4 nodes to enhance frontside SubGraph usability and organization: a Note, a Divider, an adjustable Spacer, and a Title. Note and Title text is multiline and editable in place inside the subgraph and on the subgraph front, with font size, bold, text color and note background color pickers. Compatible with the current ComfyUI frontend and backwards compatible with older versions.
Video nodes for comfyui-pytorch360convert
Apply 2D transformations to images and masks within ComfyUI. Zoom, position, scale, flip, rotate, squash and stretch the input content. Tile images to create patterns (supports alpha channels). Fill options for managing canvas areas exposed by transformations. Apply masks to RGB images and invert mask inputs or outputs. No extra dependencies.
Hangeul: regional generation and detailing for Anima, SDXL, Krea 2, MOSS and FLUX. Experimental RS Workspace app with automatic model routing, wildcards, character LoRAs, editable masks, 3D posing, camera control, image references and progressive sampling; SD1.5 is unstable.
Chain multiple images as reference conditionings for Flux Klein, Kontext etc. with one node
Expressive text-to-speech with zero-shot voice cloning. Ships with preset voices covering accents, ages, and emotions. Runs on 8GB VRAM.
Focused local llama-server lifecycle, text, vision, and structured generation nodes for ComfyUI
A ComfyUI custom node for 3D camera angle control with interactive viewport to adjust camera angles and output formatted prompt strings and processed images for multi-angle image generation. (Description by CC)
A powerful LoRA manager for ComfyUI with preview, tagging, and multi-node distribution support.
MiniMax H3 video generation on Apple Silicon, powered by the h3.c Metal engine.
[JoyAI-Image]: Awakening Spatial Intelligence in Unified Multimodal Understanding and Generation
SolarWM:Open Data and Scalable Training for Long-Horizon Video World Models,use this node to run minimaxH3
ComfyUI Video Anti-Aliasing Pack
DAVE (DC Attenuation for diVersity Enhancement) as a drop-in MODEL patch for Anima / Cosmos-predict2 DiT: recover same-prompt sample diversity by attenuating the DC component of early-mid blocks over the first tau fraction of steps. Self-contained — bundles the shipped block-pool mask, pulls no training-repo code.
3D editor node — build a blockout scene (primitives, GLB, FBX with animation, gaussian splats), record a keyframed camera path, and feed the rendered mp4 + image + prompt into Seedance / Nano Banana / Flux / any downstream model.
ComfyUI nodes for identity-preserving 2D character pose transfer (sprite A → sprite B).
ComfyUI nodes for previewing, round-tripping, and pasteback editing 360 ERP photos.
Three custom ComfyUI nodes: Empty Image RGB, Text Line Break, and Random Chars (Append) for convenient utilities. (Description by CC)
Native ComfyUI V3 nodes for AuK speech generation and editing.
A hacky implementation of Simswap based on [a/Comfyui ReActor Node 0.5.1](https://github.com/Gourieff/comfyui-reactor-node) and [a/Simswap](https://github.com/neuralchen/SimSwap).
ComfyUI nodes for merging, testing and more. NOTE: SDNext Merge, VAE Merge, MBW Layers, Repeat VAE, Quantization.
Controllable random NSFW natural-language prompt generator for ComfyUI.
Marigold V2 in ComfyUI: single-step depth, surface normals, and albedo from one image.
A structured prompt-builder node for Krea-2 / K2: fill scene, N characters, composition, lighting, and style fields and assemble them into Krea2-native flowing prose that feeds straight into Text Encode (Krea2).
Extensible content viewer for ComfyUI: HTML, Markdown (LaTeX, Mermaid, etc), Code, Images, and more
All-in-One Face Fix KSampler for ComfyUI with YOLO detection and SAM segmentation
Powerful REST API nodes for ComfyUI that enable seamless HTTP/REST integration into your workflows.
Cyberdolphin Suite of ComfyUI nodes for wiring up things.
Assist in determining reasonable input values for tiled_width/tiled_height in the UltimateSDUpscale node to improve efficiency.
MiniMax H3 video and audio generation nodes for ComfyUI
One-click fix for deformed / broken / blurry faces in MiniMax H3 videos, single or multiple people.
Nodes:TacoLatent, TacoAnimatedLoader, TacoImg2ImgAnimatedLoader, TacoGifMaker.
ComfyUI-ChatterboxTTS is now available in ComfyUI, Chatterbox TTS is the first production-grade open-source TTS model.
ComfyUI-Step1X-3D is now available in ComfyUI, delivering high-fidelity 3D asset generation with consistent geometry-texture alignment. It supports multi-style outputs: cartoon, sketch, and photorealistic.
ComfyUI sidebar extension — brings back the queue panel with image previews
A lightweight and high-speed ComfyUI custom node for generating image captions using BLIP models. Optimized for both GPU and CPU environments to deliver fast and efficient caption generation.
ComfyUI-KokoroTTS is a powerful text-to-speech node for ComfyUI. It leverages the KokoroTTS to convert text into natural-sounding speech, supporting multiple languages.
High-resolution 2K video regeneration and refinement nodes for MiniMax H3 in ComfyUI using Open-H3-Regenerate-2K LoRA, Sharp AdaLN curve branches, and audio-locked latent refinement.
Detect and resolve models missing from the active ComfyUI workflow graph, with multi-source resolution (Civitai / Hugging Face) and checksum-verified downloads.
ComfyUI任意模型路径修复工具 | 智能模型路径修复引擎。告别 红框 报错,一键修复 ComfyUI 模型加载器路径错误。
A powerful ComfyUI custom node for API-based image editing with multiple provider support
a custom node for [a/FireRedTTS](https://github.com/FireRedTeam/FireRedTTS)
Lightweight CPU-based Text-to-Speech nodes for ComfyUI. Fast, efficient TTS running at 6x real-time on CPU without GPU. Supports voice cloning and 8 built-in voices.
Krea 2, 2.42x faster than BF16 on an RTX 3090 (up to 2.90x), at a third the size. Native INT8/W4A4 quantization with an SVDQuant low-rank branch, no calibration dataset needed. Turbo and base.
Custom ComfyUI nodes by axces2000 — Audio Loader, Resolution Master, and more
Nodes for Iris-3B, a 3B pixel-space text-to-image model, plus its 4x upscaler and depth model with a 3D view. Simple, advanced and 4-step Turbo samplers, image to image, auto model download, low VRAM.
Nodes:Apply Invisible Watermark, Extract Watermark. Adds up to 12 characters encoded into an image that can be extracted.
A ComfyUI node for transforming images into descriptive text using templated visual question answering. Leverages Hugging Face's VQA models with transformers
Cicada AI Nodes for ComfyUI.
A small collection of nodes intended for use with Lodestone Rock's Chroma model, for ComfyUI.
Import soda-storyboard bundles or MOSS layout JSON, preview the page and prompt, and render manga pages with MOSS-Image: slanted panel cutting, Chinese lettering and panel fixes.
ComfyGit Manager - Node for managing comfygit environments in ComfyUI
Example node communicating between ComfyUI Javascript and Python.
A video timeline studio inside ComfyUI: cut clips, generate takes with your own workflows locally or on Comfy Cloud, and choose the take that fits each shot.