ComfyUI Custom Extensions
Browse 9,446 ComfyUI extensions and 73,215 custom nodes.
100% of code taken from [a/https://gist.github.com/neggles/ecb6327251a9e274428d07636c727eb9](https://gist.github.com/neggles/ecb6327251a9e274428d07636c727eb9).
Connect your ComfyUI workflows directly to Autodesk Flow (formerly ShotGrid). Browse projects, select shots, assign tasks, and publish versions—all from within ComfyUI.
A custom node for ComfyUI that allows you to save generated images while completely removing all ComfyUI and AI metadata and injecting fake EXIF metadata. (Description by CC)
A custom node for ComfyUI that allows temporal switching between prompts.
concatenate all images in floders, concatenate caption txt files for trainning loras.
Full-pipeline virtual try-on and pose transfer ComfyUI custom node, built on Meta's Leffa.
ComfyUI node for converting images to IRC art blocks
Nodes:BrevImage. ComfyUI Load Image From URL
This is a plugin for ComfyUI that adds new Util Nodes and Nodes for easier image creation and sharing.
Switch nodes: pick from up to 10 named inputs of any type from one dropdown (only the picked one runs, the rest are lazily skipped), and a resolution switch with aspect-ratio/megapixel, manual width/height, or from-image modes.
A ComfyUI extension for using PixelDojo's Flux API to generate high-quality images directly within ComfyUI workflows.
A custom node collection for ComfyUI designed to handle advanced media operations including switching between image and video inputs, face swapping, color matching, and video frame interpolation.
This script adds an automatic batch runner UI to ComfyUI. The underscore in ComfyUI_AutoBatchRunner is a naming error. Please use ComfyUI-AutoBatchRunner.
Nodes:LoraWithMetadata, TypecasterImage.
Custom node to reboot ComfyUI server
Calculator
HunyuanPortrait enthusiast level node pack for experimental and educational exploration
Collapsible Nested Menus and Save Workflow as PNG
Pick an LLM role (system prompt) from a dropdown and output it as a STRING for an LLM node's system_prompt. Modes: fixed, increment, or random over a start..end slice, with index/seed control_after_generate to step roles across a batch. Roles are plain Markdown files in roles/; add a .md file and restart to add a role. Node name: Set Role. Located under Add Node > LLM.
Nodes: CLIPTextEncodeSDXL-Multi-IO, CLIPTextEncodeSDXL-Pipe, Empty Latent Image from Aspect-Ratio, Random Find and Replace.
ComfyUI nodes for Groq API - chat, vision, audio, tool use, and batch processing
Prompt optimization, local AI forensics, and image provenance for ComfyUI. Includes the Mutant Prompt Optimizer (Word Weight methodology), Mutant Forensic Lab (local FFT/ELA/metadata detection), and Mutant Signature (PNG provenance injection).
Prompt composition, image generation, LoRA management, upscaling, detailing, and all-in-one workflows for ComfyUI.
Experimental unified-memory-aware DGX Spark / GB10 loader nodes for ComfyUI
This custom node allows users to selectively choose what to add to the generated image's metadata.
Custom ComfyUI nodes for metadata-first prompt, generation, image, and save workflows.
Nodes: OTX Multiple Values, OTX KSampler Feeder. This extension provides custom nodes for ComfyUI created for personal projects. Made available for reference. Nodes may be updated or changed intermittently or not at all. Review & test before use.
Prompt Builder, List Pick and String Line Count: text and list utility nodes for building and iterating prompts.
Generate random prompts easily for FMJ using TXT files (1 prompt per file).
Image saving with complete metadata (prompts, seed, software versions) + smart loading.
A per channel implementation of the scaled CFG from this paper: [a/https://arxiv.org/abs/2503.18886](https://arxiv.org/abs/2503.18886)
Perpendicular CFG for reducing oversaturation issues with high guidance scale values.
A custom node for randomized prompt generation with __wildcard__ expansion, subfolder filtering, prefix/suffix support, and seed-based reproducibility.
Improved version of Kijai's Ideogram 4 Prompt Builder with workflow QoL enhancements, character/object stores, and multi-select editing.
MiniMax H3 in one node: text, image, first + last, end frame or references to video with native stereo sound. Auto keyframe header, mid keyframe, audio drive, Turbo, one-click model downloads, player, history and shot chaining. by Enzino.
A powerful batch image generation node for ComfyUI using the YK-Banana2 Pro API to generate multiple images in parallel with reference image support. (Description by CC)
Nodes:CharacterManagerNode, FilmGrain, FlipFlopperSameArch
Custom node for generating videos using ByteDance Seedance 2.0 API with support for person and background reference images.
A ComfyUI sampler that boosts detail only in latent regions that are already dense at the moment the schedule activates.
Group wildcards together for easier character management
A bridge for ComfyUI to seamlessly integrate with LM Studio's capabilities, enabling advanced image-to-text and text generation workflows using local models.
ComfyUI custom node for structured prompt analysis and replacement
A ComfyUI custom node package for batch image processing with filename preservation.
Extract comprehensive scene properties from 360-degree equirectangular panoramas, including depth, normals, and PBR materials, using OmniX adapters with Flux.
Professional ZMongo Legal Operations suite for ComfyUI
A ComfyUI custom node for creating mindmaps from markdown
Dynamic prompt nodes to fully support wildcards
Model-friendly cinematic workflow nodes for creating, refining, and exporting AI video in ComfyUI.
Implementation of MoMo interpolation for ComfyUI
Nodes:ImageRequestNode
A ComfyUI custom node to load images from a ZIP archive as modern batched IMAGE and MASK objects.
This is a ComfyUI custom node that automatically detects trigger words from text prompts and applies the corresponding LoRA models.
A professional suite of utility nodes for ComfyUI, featuring forensic logging, massive wildcard control, universal noise injection, and persistent counters.
Rearrange & reorder ComfyUI subgraph inputs, outputs and promoted widgets by drag and drop. Every instance updates, links stay connected and widget values stay put.
An optimized storyboard-based latent generator for Wan 2.2 with low VRAM optimization and storyboard logic for controlled video composition.
A custom ComfyUI node that sends images to OpenAI-compatible vision APIs and returns the LLM's text response for image analysis and captioning.
This plugin includes two core nodes designed to handle proportion optimization tasks of varying complexity
A comprehensive toolset of ComfyUI custom nodes for latent generation, image processing, and workflow utilities
Shift Clock Sampler: dual-clock Euler/res_multistep for MiniMax-H3 video and audio at any step count.
Nodes for video tagger.
A ComfyUI node that provides access to OpenAI's image generation and editing capabilities, including support for gpt-image-1 model with both OpenAI and Azure OpenAI providers.
ComfyUI custom node for granular Q/K/V/O attention boosting in HuMo models. Provides 12 independent controls for fine-tuning audio-driven motion response.
ComfyUI-OllamaOmni — Ollama nodes for text, vision, audio, and video. Forked and extended from comfyui-ollama by stavsap.
A ComfyUI custom node that computes MSE, PSNR, SSIM, and Perceptual Hash Hamming Distance in a single execution to provide quantitative image quality feedback for generation workflows. (Description by CC)
The CS Transform node is a custom node for ComfyUI that applies a series of transformations to an input image and mask. The transformations include scaling, rotation, and translation, all centered around a specified pivot point. The node ensures that the transformed image is properly accommodated within a canvas, which can be expanded if needed.
ComfyUI custom nodes for sending images, videos, text, and audio to Slack, plus an optional Slack-to-ComfyUI workflow trigger.
DalleImageNodes is a custom extension for ComfyUI that integrates OpenAI's DALL·E 3 API for: Image generation, Inpainting (image editing), Image variation. This project supports the latest OpenAI Python SDK (v1.x) and automatically handles image resizing and format requirements (RGBA, fixed sizes) based on the examples from the offical Dall-E website.
A collection of custom nodes for ComfyUI focused on aspect ratio management and other utilities.
A set of utility nodes for ComfyUI focused on scripting. Includes an enhanced UUID generator node.
Quality-of-life nodes for big ComfyUI libraries and batch production runs. Workflows+ replaces the workflows sidebar with a virtualized browser that stays fast at 8,000+ files — search by name or by what's inside the JSON, then pin, rename, move, duplicate or trash workflows and whole folders without leaving the panel. Batch Prompter turns any list into a batch, queueing one run per line of a text block: prompts, filenames, LoRA names, seed labels. Model Path Auto-Fix repoints models and LoRAs saved under folder names your install doesn't use, with an override list for renamed or requantized files. Also included: a streaming TensorRT upscale + FILM interpolation node that encodes long videos in bounded RAM, and Text Preview and Edit, which writes the resolved prompt into your output's metadata so you can read it back later. See the README for the full node list.
Auto aspect-ratio padding/unpadding for Qwen Image Edit and similar diffusion pipelines
Required self-hosted ComfyUI extension for ComfierUI remote clients.
ComfyUI custom-node package that fetches node descriptors from Remote Node Protocol v1 (RNP/1) servers and exposes them as native V3 nodes.
ComfyUI custom node for JPEG/PNG compression with quality, scale, and performance reporting
NODES: Daisy-Chainable String Constant, Two-Way Switch, Three-Way Switch, Preview Imag Miscellaneous custom nodes for ComfyUI
Example of ComfyUI Node Localization
A prompt-generator for ComfyUI, utilizing the power of a language model to turn a provided text-to-image prompt into a more detailed and improved prompt.
A ComfyUI custom node for vision + text analysis using latest GPT-5.2, GPT-4 and O1 models with direct OpenAI API key input, system prompt, temperature, max tokens, and multi-image support.
Custom node for ComfyUI that integrates Google's Imagen 4 and Imagen 3 generative models via the Gemini API for high-quality image generation.
A ComfyUI custom node that automatically calculates optimal output dimensions for Google's Nano Banana image editing model, supporting 22 aspect ratio buckets and ensuring pixel-perfect outputs without shifting or cropping.
Save layered TIFFs and flat images from ComfyUI with XMP metadata burned into the file — workflow graph, prompt, model SHA256 hashes, and arbitrary JSON embedded per-file.
ComfyUI node that sends text + a system prompt to an OpenAI-compatible endpoint and returns the reply with <think> blocks stripped.
Flexible ComfyUI node that simplifies prompt building through organized dropdown categories for tag-based models.
This ComfyUI custom node processes SVG (Scalable Vector Graphics) images to ensure the artwork consistently fills its canvas, is visually centered, and optionally has a margin applied. It's designed to provide more accurate results for complex vector graphics, such as those generated by vector tracing tools, where simple geometric bounding boxes may not align with perceived visual extents.
Unified nodes pack primarily for Image to Image generation with a toggleable Text to Image option for easy switch back. CivitAI and local image inspection.
CWK Checkpoints Preset Manager — A ComfyUI custom node that combines a full model manager with a per-checkpoint preset system. Manage all your models from a visual panel with CivitAI metadatas, favorites, and update. Save presets (sampler, scheduler, CFG, steps, clipskip, resolution, RNG) per model.
Adjust whole comma-separated prompt tags with Ctrl+Up/Down in ComfyUI.
Example of a Combo widget changing Node appearance
some tiny nodes for better QoL
Creditas' Ad Creation Toolkit for ComfyUI
Creditas' Ad Creation Toolkit for ComfyUI
ComfyUI custom nodes for FBX loading and offscreen rendering.
A standards-compliant ComfyUI custom-node starting point with backend nodes, subgraph blueprints, and a React/TypeScript frontend extension.
A wildcard prompt node for ComfyUI with Jinja2 templating, YAML wildcard support, negative prompt resolution, and hot-reload — no server restart needed when you update your wildcard files.
Camera and photography style presets for ComfyUI, with JSON libraries, dynamic dropdowns and custom uploads. By Cyberdelia AI Lab.
Prompt cleanup for ComfyUI — port of sd-webui-prompt-format. Dedupe tags, fix bracket/comma spacing, swap underscores. Button-driven, not realtime. By Cyberdelia AI Lab.