ComfyUI Custom Extensions
Browse 9,446 ComfyUI extensions and 73,215 custom nodes.
A preset-driven MOSS-Image sampler and 128-channel empty latent node for ComfyUI, with Turbo and full-model presets, constant CFG, negative-branch layer skipping and direct manual tuning.
A collection of custom nodes for ComfyUI by Arkl1te
Provides a custom node to load config for sampler nodes from a yaml file.
ComfyUI image loader node that reads AI image metadata and traces real prompt conditioning.
ComfyUI custom nodes for interacting with the Gemini api for image and video generation prompting.
ComfyUI custom nodes for interacting with the Gemini api for image and video generation prompting.
A ComfyUI custom node for writing and inspecting prompts, with MANUAL/LLM status and last-input history.
A model-agnostic captioning framework for ComfyUI designed to generate cleaner, more consistent, auditable captions for LoRA dataset preparation through structured comparison, repeated observation, and consensus-based refinement.
Integrates LM Studio's CLI to perform Vision Language Model inference locally in ComfyUI with advanced memory management and anti-OOM protection.
A set of node for ComfyUI to create an influencer
Grok Vision Analyze — send an image (tensor or URL) to xAI's Grok vision API and get a text response back.
A custom node for ComfyUI that allows you to inspect any input data with detailed metadata and visual preview.
A simple custom node for ComfyUI that allows you to quickly select resolution presets from a dropdown menu.
Easily replace variables like *obj or *style in your prompts using templates.
Bypass or mute entire node groups with a single click. Includes a Block Bypass Switch node, Group Bypass right-click menu, and a floating Group Bypass Panel.
Save images directly to URL, e.g., OSS. Just input the url in the text box!
Image preview and manual or automatic saving with generation metadata and optional workflow.
DARASK custom nodes: folder-driven image queue + EXIF/PNGinfo-driven model/LoRA/prompt restoration for upscale & hires-fix workflows, plus an aspect-ratio-annotated empty-latent preset.
A ComfyUI custom node for Google Vertex AI / Gemini image generation with workflow-level parallel execution.
A custom node for ComfyUI that allows you to interact with Nano Banana's Vertex API proxy, providing access to Gemini models with dynamic image inputs.
ComfyUI custom nodes for preventing OOM crashes with real-time VRAM monitoring, automatic model unloading, LRU cache, and memory budget gates.
This node calculates a contrasting complementary color based on an input RGB color. The goal is to ensure visibility and contrast when overlaying text, UI elements, or graphical components against a given background color.
This project generates an image from Markdown text using imgkit and wkhtmltoimage. It automatically scales the text to fit within the specified image dimensions.
Discord Rich Presence integration for ComfyUI - shows generation status in Discord
A custom node pack for ComfyUI that enables creative manipulation and 'bending' of neural network models. Perform various operations on loaded model checkpoints to create unique and experimental effects.
Multi-port image save node: wire up to 12 IMAGE inputs (40 with the dynamic-port version) and it saves whichever port is still alive, falling back a level whenever an upstream stage is muted or bypassed. Also strips prompt, workflow and EXIF metadata. Standard library plus the torch/Pillow that ship with ComfyUI.
A ComfyUI node implementation of the Dynamic Consistency Weighting (DCW) technique that enables high-quality image generation with drastically reduced inference steps.
Smart switch node for ComfyUI with dropdown-based input selection and INT override for cascading workflows. Route any data type through a named dropdown instead of moving noodles.
Automatically resizes masks to match image dimensions for seamless inpainting workflows.
Compare two texts with GitHub/GitLab-style diff highlighting in ComfyUI
Official EvoLink nodes — 56 nodes, one node per model variant: Nano Banana 2/Pro/Lite, Seedream 5.0, GPT Image 2, Qwen Image 3.0, Z-Image Turbo, Krea 2, Grok Imagine, Sora 2/Pro, Veo 3.1, Kling V3/Turbo/O3 (ref/edit/motion), Seedance 2.5/2.0 (ref/edit/extend), Wan 3.0, Hailuo 2.3, Gemini Omni Flash, Topaz Upscale. Chain generate→edit→upscale via result_urls. No local GPU, no third-party image hosting, async & cancellable, API key stays out of workflow JSON. 国内直连,中文支持。
Save toolkit for audio, image, and video: MP3 audio export with VBR/CBR options, multi-format image saving with workflow/thumbnail metadata, and video + frame encoding (MP4/MKV/WEBM/MOV) — all sharing whitelist-safe paths and rich placeholder templating.
Lightweight node to number text divisions (one per newline) and select single/multiple divisions by index with an optional separator
Lightweight node that generates a string of numbers between a start and end value. Supports positive and negative steps, configurable separators, and inclusive/exclusive end values.
VoxCPM custom nodes for ComfyUI - High-quality Text-to-Speech with Audio Reference
Maintenance fork of ChenDarYen/ComfyUI-NAG (Normalized Attention Guidance) with compatibility patches for ComfyUI v0.21+.
ComfyUI outfit captioning nodes for virtual try-on prompts, backed by NVIDIA NIM and Google Gemini.
Organized image saving for ComfyUI with model-aware folders and template variables.
Training-free Temporal State Transport (TST) temporal attention correction for MiniMax H3 video generation in ComfyUI.
ComfyUI custom node for separating instrumental and vocal stems using the bundled MDX model without external UVR dependencies.
ComfyUI custom nodes for Hi3D Image-to-3D API: generate 3D models (GLB) from images.
ComfyUI custom node package for Hitem3D Image-to-3D API. Quickly generate high-quality 3D models from images using Hitem3D AI services.
Complete 25-node Anima ecosystem suite for ComfyUI: Native Diffusers-aware Model/Qwen3-CLIP/16ch-VAE Loaders, Multi-Format ConvRot INT8/INT4 Model Saver, LoRA Extractor, Block Weight & SVD Resizer, Offline Danbooru Artist/Character Picker, Smart Tiled VAE Decode, Corner Scribble Shield, Block-level Model/CLIP/LoRA Mergers, TIES/DARE/Model Stock/SLERP Multi-Model Merger, QR-SVD Exporter, Prompt Studio & Aesthetic Slop Guard, NAG CFG-1.0 Attention Guidance, 16ch Latent/VAE Guard, Universal Diffusers LoRA Remapper, Single/Dual/Triple-Pass Anima Rectified Flow KSamplers, Image Saver Sampler Metadata, and Centralized Suite Telemetry & Flight Auditor.
A ComfyUI node implementing cascade object detection system (bbox/segm) using YOLO models with multi-stage processing. (Description by CC)
Turn pixel data into ascii art
Dexogen's personal collection of custom nodes for ComfyUI, inspired by Pixaroma.
A custom node for ComfyUI that creates looping animations from image sequences.
CONDITIONING Color Bias Node for ComfyUI
A lightweight ComfyUI custom node package for final image color correction, designed for quick exposure, contrast, saturation, temperature, tint, shadow, and highlight adjustments.
Experimental ComfyUI nodes for trellis2-apple, using its MLX backend on Apple Silicon for image-to-3D generation.
Custom ComfyUI nodes for a tarot card reader workflow.
A ComfyUI custom node that enhances and rewrites prompts using OpenAI-compatible or native Ollama APIs with optional vision-enabled image input support.
Yet another llama_cpp node. GGUF-focused vision-language captioning via llama-server, with automatic server lifecycle management and VRAM-safe handoff.
Lightweight NVIDIA hardware monitor for ComfyUI. No node conflicts, pure display widget.
Castcut's checks inside ComfyUI: in-job pose score and best-of-two pick, matte repair, and HTTP routes that answer face, pose and two-person pose checks on the CPU without waiting for the render queue, plus input-folder clean-up and health for the Castcut app.
Making ComfyUI more comfortable.
Powerful ComfyUI nodes for Supabase integration - Database, Storage, Realtime, Auth, Edge Functions
This node pack offers various ways to build json strings that is ment to be used with a LLM to generate Prompts. It also have a node that can load different prompts so that a LLM can use the json string to make the Image Prompt.
MOSS-Image prompt with up to 4 reference images read by the text encoder, each with its own strength slider, plus one for the prompt's words. No VAE or LoRA needed.
Aspect-ratio calculator node for ComfyUI — preset + custom ratios, base-resolution scaling, dimensions snapped to multiples of 8, with a live web preview.
Validated MiniMax H3 accepted-path BlockCache with finite-safe FL2VA and Ref2VA probe scoring.
Scrubable inline WebM video preview for ComfyUI image batches, with optional audio. Encodes VP9 into the temp directory and plays back in the node — nothing is saved.
Automatically tests the impact of each phrase in a prompt by generating images with one phrase omitted at a time.
A ComfyUI custom node that replaces the stock Load CLIP node with one that stores text encoder weights as FP8 buffers on the GPU, nearly halving persistent VRAM consumption for large models while preserving full inference correctness.
This is a node to generate new image that combine 2 objects from different scene.
NODES: DY Image Quantize, DY Image Cluster, DY Image Palette, DY Image Masks, Image List to Grid, DY Image Scaler, DY Random Lines, DY Adaptive Color Lines, DY Adaptive Color Circles, DY Adaptive Color Rectangles, DY Binary Pattern Stamper
Google Gemini image generation node for ComfyUI supporting up to 6 input images with customizable prompts and parameters. (Description by CC)
Official EcoHash API nodes for public and dedicated models: image and video generation, image editing, LLM and vision tools, TTS and STT. One EcoHash API key, no local model weights required.
Custom node for ComfyUI to show values of primitives (str, float, int, or bool).
ComfyUI custom nodes for 9D Human Pose Estimation, integrating the [a/HSyn9DHPE](https://github.com/edoardocompagnucci/HSyn9DHPE) project.
Using the ollama gpt-oss:20b model, a prompt is generated. The user enters the desired theme into a custom node and runs it, and the model outputs a Stable Diffusion prompt. This is useful when you want the model to handle prompt creation for you.
Professional noise generation nodes for ComfyUI
Tiled sampling and latent caching for MiniMax H3 video generation in ComfyUI
Paste an AI edit (Qwen-Image-Edit, FLUX Kontext / FLUX.2 Klein, MiniMax H3 Image, GPT-Image, Nano Banana) back into the original without drift: aligns geometry and colour, finds the changed area, keeps every other pixel bit-exact.
A collection of custom nodes for ComfyUI designed to enhance image processing workflows. Especially useful for high-resolution rendering, complex inpainting, tiling, and batch manipulation. This allows you to perform processing that would otherwise exceed your VRAM limits.
Custom node for ComfyUI that splits a multiline prompt into 5 independent CLIP conditionings.
A ComfyUI custom node that generates structured multi-character scene prompts with spatial layout control, supporting composition and scenario customization for image generation models.
API friendly save text to file.
ComfyUI diagnostic utility for ROCm/HIP Comfy Kitchen builds: It reports either detailed or summarized Comfy Kitchen kernel dispatch info, and other insightful aspects, such as hardware/stack performance. The custom node has the ability to persistively store the reports, just by simply being wired anywhere within functional ComfyUI workflows (which includes placement even after any 'Save' Image/Video/Audio/Latent/etc node that has an output to hook into). Project also features flag/env var option to display summary or detailed reports via terminal, without involvement of the custom node. Intentionally very lightweight, utilizing just psutil, WDDM, and ultimately optimized and intended for AMD GPU users on Windows.
ComfyUI custom node to rotate OpenPose figures around their torso
Custom nodes for ComfyUI: Drag & drop image paste, interactive UI crop, WAN I2V bucket resolution, and GIMM-VFI frame interpolation.
Converts normal maps to depth maps for use in ComfyUI. (Description by CC)
EmpirioLabs AI nodes for ComfyUI: generate images and videos, chat with text and vision models, and turn text into speech with models from the EmpirioLabs AI catalog.
Show markdown notes of checkpoints and loras.
ComfyUI nodes for OpenFormosa BlueMagpie-TTS
ComfyUI local-loader custom node for bosonai/higgs-audio-v3-tts-4b
ComfyUI custom node for running baidu/Unlimited-OCR locally with Transformers.
VideoChat3-4B visual video understanding nodes for ComfyUI
Commercial-friendly Whisper Large V3 transcription nodes for ComfyUI.
ComfyUI custom nodes by enigmatice: mask utilities and analog video effects
Small, dependency-light ComfyUI utility nodes from Enviral Design.
Native pixel-space Iris-3B inference for ComfyUI
Automatically refresh file-backed node choices using native filesystem notifications and selective node-definition updates.
A ComfyUI custom node for translating text prompts using local GGUF models via llama-cpp-python, supporting English or Chinese without internet.
A ComfyUI custom node for NSFW (Not Safe For Work) content detection using YOLO models with optional blur functionality.
ComfyUI custom nodes for Alice T2V 14B MoE text-to-video generation
Agentic web research nodes for ComfyUI — grounded prompt generation and VL reference description
ComfyUI custom node suite combining resolution/prompt nodes and Booru Roulette.
Sweep a folder of LoRA training epochs: pick a step range + stride and sample one image per epoch from a single queue.
Temporary, experimental ComfyUI nodes implementing AbleJones's mixed-timestep generation-mask fix for MiniMax H3 without modifying ComfyUI core files.