ComfyUI Custom Extensions
Browse 9,446 ComfyUI extensions and 73,215 custom nodes.
Intended to be used with SwarmUI-rtx-upscale; provides RTX support with simplified enum handling.
Strip metadata & workflow info from ComfyUI output images/videos (PNG/JPEG/MP4/WebM/GIF/WebP, lossless). 无损去除图片/视频中的提示词、工作流与EXIF信息。
Provides recommended resolution presets specifically. Select a preset from the dropdown to easily output the corresponding width and height values for use with nodes like Empty Latent Image.
Custom node that displays the intermediate images generated at each denoising step during sampling. (Description by CC)
Some additional audio utilites for use on top of Sample Diffusion ComfyUI Extension
parakeet-tdt-0.6b-v2: Automatic speech recognition (ASR) model designed for high-quality English transcription, featuring support for punctuation, capitalization, and accurate timestamp prediction.
Interactive editor for drawing animated, keyframed masks on a video/image batch in ComfyUI — circle, rectangle and bézier shapes with automatic keyframing, feathering and invert.
Scoped wildcard selector for ComfyUI. Pick a theme and categories from a bundled 5,200+ entry library with metatype-driven skin tone; get a finished prompt string. No wildcard syntax required.
AMD Unified Memory TeaCache for LTX2 (LTXAV). Lean single-file caching for 48-layer dual-stream transformer. Caches both video and audio residuals. No CPU offload - designed for AMD APUs with unified memory. ~35% speedup on cached denoising steps.
EzFlex preset suite for ComfyUI: models combo loader, resolution/latent selector, node switch master & group presets, parameter presets, preview any, prompt helper, media loader/output.
This model was fine-tuned on Flux.1-Dev to with reflection tuning to serve as a corrector for self-refinement framework introduced in From Reflection to Perfection: Scaling Inference-Time Optimization for Text-to-Image Diffusion Models via Reflection Tuning.
ComfyUI nodes for processing SVG and generating G-code using vpype
Visual enhancement extension that adds animated arrows to links on hover and highlights Set/Get node pairs.
Another comfy implementation for the short video generation project hpcaitech/Open-Sora, supporting latest V2 and V3 models as well as image to video functions, etc.
ComfyUI custom node for BriaAI Fibo Edit image model.
Fixed version of the original [a/MagicQuill](https://github.com/magic-quill/ComfyUI_MagicQuill) node. Required nodes: ComfyUI-Brushnet and ComfyUI Controlnet AUX
Official BudgetPixel nodes for ComfyUI — generate images, video, music and sound effects with 60+ hosted models (FLUX, Seedream, Kling, Seedance, Wan, and more) using your BudgetPixel API key.
Generate random prompts easily for FMJ.
ComfyUI nodes for BytePlus ModelArk (Seedream, Seedance, Seed/DeepSeek/GLM LLMs, Hyper3D and Hitem3D 3D models), Seed Speech (Seed Audio, TTS, ASR), Vision AI (OmniHuman avatar video) and AI MediaKit (video and image enhancement), billed to your own BytePlus account. Docs: https://byteplus-sa.github.io/ComfyUI-BytePlus-ModelArk/
A custom node implementation for ComfyUI that integrates with Together.ai's FLUX image generation models. This project is inspired by and adapted from [a/ComfyUI-FLUX-BFL-API](https://github.com/gelasdev/ComfyUI-FLUX-BFL-API) to work with the Together.ai API.
ComfyUI sidebar extension for browsing, searching, and managing generated images and input assets. Scans input, output, and custom directories; indexes PNG/WebP metadata into SQLite for fast filtering by model, LoRA, sampler, prompt, and more; supports favorites, NSFW filtering, workflow loading, and an integrated fullscreen image viewer.
Signal Processing Nodes For ComfyUI
ComfyUI nodes for loading images and drawing polygon masks interactively on them with a polygon lasso style tool
A custom node pack for ComfyUI that provides an intuitive prompt drafting, saving, loading and management system with wildcard support.
A ComfyUI custom node pack that resizes images to an exact target canvas size with full control over scaling, cropping, and placement.
Fork of [a/sd_webui_stealth_pnginfo](https://github.com/ashen-sensored/sd_webui_stealth_pnginfo) with ComfyUI support.
Nodes:Simple Google Translator, Simple Resolution Solver. This is a small set of simple nodes that help your workflow on ComfyUI.
🤖 ControlNet-aux-style preprocessor nodes for ComfyUI — Depth/Normal/Soft-Edge/Line/Pose map estimation, ported as real architecture code (not a thin wrapper) from Fannovel16/comfyui_controlnet_aux.
🚀 Picks a random IMAGE file from a directory and provides you with the IMAGE and its filepath as a STRING! This awesome tool brings a touch of randomness to your files and images, making your file management tasks a breeze.
Make every ComfyUI image Civitai-ready. CiviScribe saves PNG, JPEG, and WebP with your prompts, generation settings, model, LoRAs, hashes, and only the resources actually used. PNG files can also carry a reloadable workflow, so your uploads keep their creation details without manual metadata work.
Used for batch extraction of prompt words.
ComfyUI nodes for PromptWrapper, Mainly for prompt word translation, Chinese translation into English or English translation into Chinese.
ComfyUI custom node for downloading models from CivitAI, HuggingFace, and custom URLs
🧩 Aspect Ratio Image Size Calculator, 🖼️ Aspect Ratio Resizer, and 📄 Markdown Link Generator for ComfyUI.
This custom node facilitates direct interaction with the Google Gemini API for advanced image editing tasks within ComfyUI with parallel request handling.
Example of using ComfyUI Toolbar to Toggle ComfyUI links on/off
Krea-2-Turbo text-to-image on Apple MLX (Apple Silicon), with stackable LoRAs. Runs the MLX pipeline in-process and returns ComfyUI IMAGE tensors.
ComfyUI custom node: batch prompt queue from multiline/text file with incremental per-run consumption and {p} templating.
A random prompt node
Anima-friendly first-pass resolution picker for ComfyUI. Outputs resolution, width, and height integers for image generation workflows.
Exact mixed-rank merging of 2–4 Krea 2 PEFT/Diffusers or Kohya LoRAs, with FP16/BF16 output, FP32 alpha, adapter validation, safe loading, and protected saving.
Save image node with dynamic paths and an 'Open Folder' button.
This is a custom node that allows you to easily call up and set image size presets. Settings can be made by editing the included config.yaml. It is almost identical to Comfyroll Studio's CR AspectRatio. I created it because I wanted to easily edit the presets.
Creates images using the NanoBanana API. To use it, you must enter your issued API key.
NODES:Mesh To Voxel, Voxel Block Saver, Voxel Viewer, Voxel Block Loader, Voxel Video Viewer, Voxel Blocks Into Voxel Video, Voxel Video Preview, Voxelize Mesh, ...
A bilingual prompt editor, library, history, translation, and AI assistant for ComfyUI
Custom ComfyUI node integrating Google's Gemini 3 Pro model for native image generation and processing, featuring multi-API key load balancing and auto-retry mechanisms. (Description by CC)
This custom node allows you to load data from one or more CSV files, then feed that data into subsequent nodes in a ComfyUI workflow.
A very simple Discord webhook integration node for ComfyUI that lets you post images and text.
GGUF-only build of Qwen3.5-Uncensored for ComfyUI. Uses llama-cpp-python instead of transformers — no transformers dependency, compatible with older ComfyUI nodes.
DeepSoch Toolkit — a growing collection of custom nodes for ComfyUI (prompt presets, image utilities, samplers, and more).
Prompt Style Selector with Preview, Multiple Selecting, Search and Local Base.
Custom nodes make easy Advanced Workflows. Focus on Image/Video and ControlNet efficiency and performances. Manipulation of Latent Space, Automatic pipeline with a bit efforts.
Bernini MLLM semantic planner for local prompt enhancement in ComfyUI — 12 task types, GGUF offline inference, zero VRAM residue
A lean ComfyUI node pack wrapping trellis-mac — the Apple Silicon port of Microsoft's TRELLIS.2 image-to-3D model for MPS-based 3D generation.
Release 1.0.2
This is a ComfyUI wrapper for Andrew DalPino's SuperCool upscaler, enabling its use directly within ComfyUI's node workflow. No extra dependencies required—just drop in the models and go.
Validated MiniMax H3 long-sequence Sol-Attn patch for AMD ROCm with conservative accepted-block routing.
A collection of nodes for detecting color in images, leveraging RGB and LAB color spaces. These nodes aim to distinguish colored images from black and white, including those with color tints.
Load data from Google Sheets, Google Docs, Excel, Word and TXT... Built-in login system for ComfyUI.
Automatic cropping of transparent areas to prevent images from being too large, while also supporting resizing to prevent image dimensions from being too large.
This node loads prompts (txt) and images (png) from a specified directory. By specifying an index, it outputs the selected file.
Segmented sampling and nested-latent caching for MiniMax H3 audio-video generation. This node is optimized for increasing output size (up to 2K), but it can also be used for longer videos — working around VRAM limits at the cost of longer render times.
A self-contained LoRA Slider node for ComfyUI designed to work alongside ComfyUI-Lora-Manager with bipolar strength sliders and batch LoRA loading.
Coherent sci-fi, fantasy, and horror scenes with independently lockable fields, cross-genre entities, and prose and JSON output. Offline, zero dependencies.
Professional color correction and skin masking nodes for ComfyUI
Find your LoRAs by what they do, not by folder name. A local vision LLM (LM Studio or Ollama, e.g. Qwen3-VL) writes a card for every LoRA in your library: its role (style, lighting, detailer, subject...), look, genre and mood, trigger words, good weights and NSFW level. You review and correct the cards in a fast browser editor that also searches, rates, de-duplicates and moves files with undo. Civitai data shows how each LoRA ranks against others of its base model, the weights people really use and the prompts they make with it. The LoRA Suggester node reads your prompt and stacks the LoRAs that fit, with per-stage weights and an NSFW gate. A ComfyUI sidebar panel refreshes the catalog. Everything runs on your machine; works standalone for Forge and A1111 too.
ComfyUI custom node: load an MP4 and output first/last frame as IMAGE.
Custom node system for ComfyUI enabling efficient prompt management with YAML-based datastore, categories, selection interface, and dynamic text interpolation for creating complex prompts.
A collection of nodes and utilities to make working with multiple models, custom parameters, and prompt styles in ComfyUI easier, faster, and more flexible. You are welcome to use and adapt them for your own workflows!
ComfyUI nodes for PU21/LogC HDR decode, EXR export, tone-map previews, color grading, and metrics.
GLHF is a ComfyUI node that facilitates seamless interaction with the GLHF chat API. Designed to enhance user experience, it supports multiple language models, web search integration, and customizable instructions, making it a powerful extension for AI-driven workflows.
All-in-one node pack for AI image generation. Collapses complex ComfyUI workflows into single nodes — Model Loader, Content Sampler, Reference Sampler, Image Edit, Inpaint, Head Mask, Upscaler, Face Swap, Prompt Batch, and more.
Meux ComfyTools is a collection of custom nodes that streamline common workflow chores in ComfyUI for the Meux asset platform.
A custom node for saving multiple images simultaneously with batch processing and resizing capabilities.
ComfyUI Hardware Monitor — Feixue Universal Monitor: real-time GPU/CPU/RAM monitoring for ComfyUI, AMD optimized.
An implementation of [spawner1145/TeaCache4Lumina2](https://github.com/spawner1145/TeaCache/blob/main/TeaCache4Lumina2/teacache_lumina2.py) for ComfyUI
Standalone content-aware tail-block cache nodes for ComfyUI MiniMax H3
Experimental tools and motion controls for LTX-Video in ComfyUI
FL PenguinVL - Vision-Language Model nodes for ComfyUI. Query images and videos with natural language using Tencent's Penguin-VL (2B/8B). Supports OCR, document analysis, visual QA, dense captioning, and more.
A progress bar beautification plugin designed for ComfyUI. It replaces the monotonous default progress bar with a vibrant and dynamic experience, complete with an animated character and rich visual effects.
Optimized PyTorch/Triton MiniMax H3 video VAE runtime for ComfyUI.
Steam-Mover-style GUI for shuffling model files between two or more registered storage locations, keeping extra_model_paths.yaml in sync.
This repository contains a custom node for ComfyUI that allows searching for a keyword in the first column of a CSV file and returning a value from a specified column in that row. The node is designed to be modular and fit within the node-based workflow of ComfyUI.
ComfyUI Google Image Search Node - Search and retrieve images from Google
ComfyUI tool node, which can be used for file/message backup, supports Telegram, Discord, Alist yike, etc./ComfyUI 推送节点,可用于文件/消息 备份,支持Telegram, Discord, Alist yike 等等
Save images without embedding workflow or prompt metadata
Advanced sampler nodes for ComfyUI based on efficiency-nodes, including live preview rendering and support for custom samplers with schedulers like flowmatch scheduler.
ComfyUI custom nodes for running L2P Z-Image 6B pixel-space generation, wrapping the public pipeline to enable direct no-VAE model inference.
ComfyUI nodes for remote LMStudio servers via OpenAI API v1-compatible endpoints.
Node to process negative wildcard tokens (<! !>) and move them from the positive prompt to the negative.
A lightweight ComfyUI node for size selection and size routing with Preset, Image and Custom modes.
Similar to the VHS video nodes, but with file info exposed on the node and much faster load and save for large files.
Snaps the menu to the corner automatically
Use the Tab key to switch between textfields
A powerful ComfyUI custom node that integrates Qwen2.5-VL and Qwen3-VL vision-language models with multi-image support. Process multiple images simultaneously with advanced AI capabilities for image understanding, comparison, and analysis.
ComfyUI_pxtool
ComfyUI node for image generation with Seedream 5 Pro via PiAPI.
Save and load Wan video latents
Bilateral Reference Network achieves SOTA result in multi Salient Object Segmentation dataset, this repo pack BiRefNet as ComfyUI nodes, and make this SOTA model easier use for everyone. NOTE: The original node was replaced with a version maintained by hieuck because it is no longer maintained.