ComfyUI Custom Extensions
Browse 9,446 ComfyUI extensions and 73,215 custom nodes.
Create reusable MiniMax H3 Image RefMod bundles from up to 15 image references using the H3 Image VAE. Requires ComfyUI-MiniMaxH3Mod for RefMod creation and loading.
Create complex compositions the FAST and EASY way.
Passthrough node that instantly unloads any running ollama model mid workflow. No API calls, just accepts plain string as a model name and executes 'ollama stop <model>' command in your system.
Agnes AI nodes for ComfyUI: image generation (text-to-image / image-to-image / multi-image compose) and video generation (text / keyframe / reference modes)
NovelAI cloud image generation for ComfyUI: text-to-image, img2img and inpainting with live Anlas cost preview.
Edit ComfyUI images in PhotoCraft from the image context menu, or pause a workflow for layered manual editing.
A highly advanced custom node for ComfyUI designed for creating consistent multi-image carousels using advanced Cross-Batch Identity Anchoring.
A ComfyUI node that fills in tokens like the date and finds and replaces text, with find and replace pairs that grow as you use them, and a passthrough so the date can be taken right before saving.
Picks new random seeds when a ComfyUI workflow is opened, for every seed set to randomize.
Makes ComfyUI's default workflow blank, so closing the last tab or starting ComfyUI with no tabs opens an empty workflow instead of the default Load Checkpoint and KSampler setup. Doesn't change any ComfyUI files, so updates don't undo it.
Reloads Dynamic Prompts wildcard files when they change, without restarting ComfyUI, and allows spaces between the number and the double colons '::' in weights.
Shows how long the current ComfyUI job has been running and an estimate of the time left.
Keeps Windows from going to sleep while ComfyUI jobs are running or queued.
Shows the previous run's text alongside the current run's text.
Save node settings, like prompts, images, models and seeds, as named presets, and switch between them with one click. Like snapshots of your workflow's settings, so one workflow can hold many prompt and image setups. Works well with Dynamic Prompts and Random Prompts to save different wildcard prompts and easily switch between them.
Adds a Save button to ComfyUI's top menu bar that turns green when there are unsaved changes.
A ComfyUI node that saves your workflow as a .json file with your images or videos, or into a compact archive. It can also save text such as prompts into the workflow, and save a backup of your Dynamic Prompts wildcard files.
Measurement-driven nodes for building face-centric stylistic LoRA training sets: score a portrait, then route it through only the processing it actually needs.
Video and image generation nodes for ComfyUI, backed by the Modelsite model catalog. One API key, a model list that updates itself, no comfy.org login.
Parameter sweeps for ComfyUI: try every combination of seeds, cfg, prompts, LoRA strengths, samplers or models in one run, with images named by their values, labelled grids and a CSV run list.
Simple ComfyUI node that determines the image size by a keyword in the prompt, optionally replacing the keyword
A simple ComfyUI node for sending a text prompt to Ollama's chat completion API and receiving a textual answer.
MPS-native augmentation nodes for ComfyUI optimized for Apple Silicon with high-performance alternatives for loaders, encoders, and samplers. (Description by CC)
Custom nodes for ComfyUI that load models from a Model Manager REST API with automatic caching and dynamic LoRA multi-loader.
This custom node for ComfyUI provides adaptive image resizing based on target pixel counts, maintaining aspect ratio and supporting different quality levels. It is useful for workflows that require images to fit within specific pixel budgets (e.g., for video, AI models, or memory constraints).
ComfyUI custom node for sending generated videos and workflow metadata to Telegram.
ComfyUI custom nodes for HTTP requests and API integrations.
Computer-vision nodes for ComfyUI: curated high-level nodes plus auto-generated wrappers for ~470 raw functions of the OpenCV (cv2) library. Built on the V3 node API.
Force-align a known lyric sheet onto audio, slice it by time, and repair noisy ASR transcripts against it.
Fetch PNG, JPEG, WebP and SVG images via URL. Has SSRF protection, size limits, SVG sanitisation; however, no formal security audit has been performed.
A single node, VideoFillerGrid Assemble, that joins source clips and (generated) transition clips that bridge them into one video, following an order list, and writes it straight to disk.
Custom utility nodes for ComfyUI.
ABC score preview: sheet music beside highlighted source, with optional playback from a user-supplied soundfont, and a score timeline with duration and sections.
Turn a still image and audio into a spectrum visualizer video with bars, LED dots or spline graphs, using FFmpeg's libraries through PyAV.
Generate continuous multi-clip video over a song you already have: the soundtrack stays fixed while each clip's video is generated. MiniMax H3 Ref2VA template included.
Video upscaling, resizing, interpolation, and configurable export: a Save Video with H.264/H.265/AV1 and FLAC/Opus/AAC choices, and post-processing (spandrel upscale models, Lanczos resize, RIFE 4.26) that streams one frame at a time and encodes once.
Dropdown selectors for ComfyUI
Model-scoped latent presets for ComfyUI
Model-neutral multi-reference timeline nodes for AI film workflows in ComfyUI
ComfyUI UI extension that automates batch prompt execution with configurable run counts and intervals, featuring keyboard shortcuts and a fixed control panel for streamlined workflow automation.
Shows real-time token count and adds a copy button to CLIPTextEncode nodes in ComfyUI.
ComfyUI nodes for Piper inference packages.
ComfyUI custom nodes for preserving masked pixels while sampling.
ComfyUI supports over lihxxx/DisPose, which generates a new video with a reference video as poses and a reference image as everything else.
Helper nodes to display last seed and prompt.
ShakaNodes: Utility tools for ComfyUI
The 'ComfyUI-Bongsang' is very useful tools for a diffusion model developer.
A small ComfyUI custom node that mirrors the controls and processing order from Art Venture's node while doing the image math with vectorized PyTorch ops.
Adds film grain to the image.
ComfyUI custom node pack for disk-backed raw IMAGE batch checkpoints in multistage workflows.
Task-aware Wan 2.2 sampling plans and sampler-specific breakout nodes for ComfyUI.
ComfyUI custom node for first-non-empty wildcard switching across 6 inputs.
ComfyUI custom nodes for scoped, graph-based string template variables.
ComfyUI custom node for selecting WAN resolution presets by aspect ratio.
Import settings between YAW Wan, Bernini, MiniMax H3, LTX and Krea workflows, with branch-aware mapping and a per-setting report.
A dependency-free ComfyUI image saving node with format selection, previews, and optional input-folder copies.
Predict and attribute ComfyUI custom-node breakage: removed comfy.* symbols, unbindable calls, and monkeypatches behind the upstream signature.
This is a custom node for ComfyUI that takes in a file path full of json's and finds the mape variable nodes in them and converts them to the kjnode set and get nodes.
Image viewer node that displays preview of images directly in the node.
Unload one or all Ollama models mid-workflow to free VRAM before image sampling. Wire your prompt through the passthrough socket so the node runs before the sampler. Node name: Ollama Unload (free VRAM). Located under Add Node > Ollama.
Compose seeded pose prompts from a curated pose and hand gesture library.
Some simple string tools to modify text and strings in ComfyUI.
A lightweight resource monitor for ComfyUI with CPU, RAM, NVIDIA GPU, VRAM and temperature readings.
Load images from files or URLs with live preview and source switching.
Save images with matching metadata files containing workflow and generation settings
Provides custom noise generation capabilities for ComfyUI workflows with configurable parameters. (Description by CC)
SDXL Long Context extends SDXL positional embeddings from 77 to 248 tokens using learned sinusoidal extrapolation method. (Description by CC)
NegPiP-style negative prompt weighting for Krea 2
Concatenate multiple text fields, including only the ones with an enabled checkbox
A focused custom-node pack for organized Brick saving and browsing with Save Brick Image, Save Brick Sequence, Save Brick Video, and Brick Browser sidebar tab.
ComfyUI custom node package for estimating and logging Partner/API Node credit usage per workflow, project, user, and partner node.
A single node that flushes PyTorch's GPU memory cache and runs garbage collection between pipeline stages, preventing OOM errors in multi-stage workflows (e.g. two-pass video generation). Accepts a LATENT input and returns it unchanged.
Grok node that describes an image and generates a T2I prompt to recreate it
Pair ComfyUI with ComfyRemote and send workflows for review
A custom node for ComfyUI that extracts text segments based on specified start and/or end marker strings. You can define multiple marker pairs to extract different segments from the same input text. The node intelligently adapts its extraction based on whether you provide a start marker, an end marker, or both for each pair.
Custom nodes for general workflow quality of life including resolutions sorted by aspect ratio, upscaling helps, and unique file names
Save images with Civitai-compatible metadata
Run ComfyUI workflows on cloud GPUs in your own Modal account, straight from your local ComfyUI. Sign in, deploy and run without a terminal.
ComfyUI nodes for palette-based color gel tinting, batch/preview helpers, and pixel-friendly saving.
A ComfyUI custom node that automatically starts a local HTTPS proxy server, enabling web applications deployed on public HTTPS pages to securely connect to your local ComfyUI instance.
ComfyUI custom node that saves images as WebP.
Comprehensive suite of utility nodes for ComfyUI including AspectToRes, ImageInfo, and TextProcessor for image processing workflows.
A ComfyUI node that automatically computes optimal width/height for supported models, aspect ratios, and generation modes with instant feedback.
Batch-render anime/illustration images across base + scenario combinations on ComfyUI.
A ComfyUI custom node to overlay frame numbers on images (e.g., from VAE Decode) with customizable prefix, zero-padding, font, position, and text outline.
NODES: Save Image (JPG), Combine Conditionings (Weighted)
ComfyUI custom nodes for Google GenMedia models (Veo, Imagen, Nano Banana, Gemini, Virtual-try-on, Lyria) with Gemini API key and Vertex AI dual auth support
ComfyUI dynamic I/O bridge for AI agents
About the comfyui image selector, image adjustment (panning, rotation, zoom), adjust image size to be a multiple of 8
Nodes:Jamworks_Login, Jamworks_Download, Shell_Command. A Simple Client for Jamworks Platform DAM Integration
ComfyUI node: Black Forest Labs FLUX image generation and editing via BFL API
ComfyUI node: pick format + resolution from organized dropdown, outputs width/height
A ComfyUI utility plugin designed to optimize the latent space for generating high-quality results. It approximates the closest size model for better generation results.
This is a simple plugin for ComfyUI that allows you to import A1111 CSV styles into ComfyUI prompts.
Nodes:Image Emoji Overlay
Context menu extension for CLIPTextEncode (sort prompt), EmptyLatentImage (sdxl size selector).
a tool for prompt collect
This custom node for ComfyUI provides integration with the BFL (Black Forest Labs) API to enhance and optimize image processing. It allows the use of Flux Pro directly within ComfyUI, offering advanced image processing capabilities.
Un nodo personalizado para ComfyUI que ajusta imágenes a diferentes relaciones de aspecto manteniendo las proporciones originales.