ComfyUI Custom Extensions
Browse 9,446 ComfyUI extensions and 73,215 custom nodes.
A collection of Nodes for personal use.
Inferway H3 video generation nodes for ComfyUI (inferway-comfy): text-to-video, first/last-frame video, reference-image video, task resume, remote cancel, task history and MiMo prompt expansion, with queue progress and an in-app API key panel. Runs inferway/minimax-h3-768p through the Inferway API.
ComfyUI custom nodes for bar censor (parallel solid-color bars) and mask merging.
A small ComfyUI helper for iterative prompt improvement with Ollama.
A crazy node that pragmatically just enhances a given prompt with various descriptions in the hope that the image quality just increase and prompting just gets easier.
Workaround that lets ComfyUI's LoRA hooks run on Dynamic VRAM instead of falling back to a slow full-load copy. 466 s -> 203 s in a tiled-upscale workflow, same images.
A ComfyUI node pack for working with multiple masks at once — spatial sorting, index picking, duplicate removal, and overlap-free expansion/feathering. Native-matched dilation and blur.
Circular TXT prompt sequences with browser upload, queue-time advancement, and used-prompt metadata.
ComfyUI node for [a/ASCII art controlnet](https://civitai.com/models/986392)
Swap a head easily and FAST, in one node: load a body and a head, press Run. Custom tuned checkpoint, optional LLM, preview with Save button, no wires.
Intentionally degrade Flux dev.1's prompt adherence to recreate the loose, random, *hallucinatory* feel of SD1.5 / SDXL.
3DA Scene: place a body pose and props into a location photo in 3D and render guides (playblast, init, strength) for image models.
Apply multiple LoRAs with matching trigger-word prompts in ComfyUI.
A custom node pack for ComfyUI that provides face semantic segmentation using the jonathandinu/face-parsing model, parsing faces into 19 distinct regions and generating precise masks for skin retouching, inpainting, makeup transfer, and more.
ComfyUI custom node set for efficient prompt management and visualization across categories like characters, art styles, clothing, and poses with dynamic control and integration. (Description by CC)
Nodes:SizeFromArray
ComfyUI custom node for local account and LDAP based authentication
Cross-pack masterwork workflow templates combining NanoBanana2 + Kling-Direct + ElevenLabs-Pro + NanoBanana-FaceSwap + Utility-MegaPack.
Comprehensive ComfyUI utility node pack with 175+ operations across 8 nodes and 15 swappable visual themes.
Custom nodes for loading and applying KQUBE (.kqube) cubic projection layers onto diffusion models (SDXL)
Experimental text encoder offloading for Apple Silicon / Mac setups.
DeployCash, a tool that converts workflows into WeChat and H5 with one click, and supports payments.
Visual comparison and retro transition nodes for ComfyUI, creating animated transitions between source and generated images with multiple wipe effects.
Edge Painter: a Load Image you can draw guide strokes on with the built-in MaskEditor (white-on-black strokes, zone mask, guide fusion). Edge Pencil: retouch a Canny/Lineart guide at the original size with a white/black pencil (X toggles). The regular MaskEditor stays untouched.
Custom nodes for ComfyUI focused on Flux workflows, metadata inspection, and reusable utilities.
Rasterize native ComfyUI SVG inputs directly to high-resolution IMAGE and MASK tensors.
int8 dp4a (sm_70/Volta) acceleration for diffusion DiTs in ComfyUI, over the fni8 kernels
Drop-in replacement for comfy-core Create Bounding Boxes with a full in-node canvas editor (obj/text tabs, grid snapping, inline label editing, color palette) that auto-rescales saved boxes when width/height are overridden by a linked input such as ResolutionSelector.
Save ComfyUI images, audio, and video directly to Eagle with metadata and previews.
Simple animated arrows traveling along workflow links (input → output)
A powerful collection of custom nodes for ComfyUI that enhance prompt handling, manipulation, and analysis for better AI image generation results.
Separate any image into editable layers with background inpainting using Image to Layers (https://www.imagetolayers.com).
A set of custom ComfyUI nodes for generating random integers within a range, adjusted to the nearest multiple of a user-defined divisor. Needlessly includes both an efficient and simple list implementation. Updates with each generation.
Run HiDream PS-SR as an isolated WSL post-processing node for ComfyUI videos
Simple Nodes: A set of nodes to simplify some tasks
Custom ComfyUI nodes for the VOID workflow with five stages: black mask, VLM prep, grey mask, quadmask, and modular PQ5 inference. (Description by CC)
Optional companion to ComfyUI-Ultimate-Auto-Sampler-Config-Grid-Testing-Suite. Distributes grid generation across multiple ComfyUI machines on a LAN (master + worker model).
A ComfyUI node for scaling pixel art using various algorithms
Flow field + particle trails custom node for ComfyUI
Optimize OOM issues based on ComfyUI-LatentSyncWrapper. [a/ShmuelRonen/ComfyUI-LatentSyncWrapper](https://github.com/ShmuelRonen/ComfyUI-LatentSyncWrapper) VideBasic Optimize OOM Plan: [a/jax-explorer/ComfyUI-VideoBasic](https://github.com/jax-explorer/ComfyUI-VideoBasic)
Nodes:FastImageListToImageBatch
Motion-aware denoising cache and batched VAE decoder for MiniMax H3 in ComfyUI
Melite Audio (Audiocore TTS/SFX/Music) — a Melite first-party ComfyUI extension pack
Melite AutoRig (SOMA weight transfer + rigged GLB persist) — a Melite first-party ComfyUI extension pack
Melite Kimodo (text-to-pose, multihmr2 + kimodo) — a Melite first-party ComfyUI extension pack
Wrap diffusers-style pipelines so ComfyUI's model_management tracks their VRAM
Melite Poser (poser server: skin/identity/presets/render) — a Melite first-party ComfyUI extension pack
Melite SFX (Audiocore sound effects) — a Melite first-party ComfyUI extension pack
Melite TTS (Audiocore speech) — a Melite first-party ComfyUI extension pack
Melite Unity (MCP-driven engine lane) — a Melite first-party ComfyUI extension pack
Melite Utility (model dirs, GLB preview, plots) — a Melite first-party ComfyUI extension pack
ComfyUI custom node pack for aggregating UCLA campus news, events, and daily briefs.
ComfyUI custom node for calculating tile dimensions in upscaling workflows
MiniMax H3 (fl2va) video+audio refinement: encodes a draft into an AV latent and builds the conditioning for a low-denoise sampler.
Usage profiles for ComfyUI: install, check and uninstall from the sidebar the models, custom nodes, workflows and pip packages of a use case, described in a profiles.json.
ComfyUI custom nodes for Octopozt Ad System with identity lock, product preservation and brand style analysis.
One finished prompt and one seed per plate. Pair with Start Loop and Get Item From List.
A custom ComfyUI node designed to inject programmatic quality-assurance guardrails into generative AI pipelines.
Peek at any node output without adding preview nodes — Alt+click an output socket for an instant floating preview (cache-aware, zero graph clutter)
Direct OpenAI GPT Image 2.6 and 2.5 (Flare / Sunburst), GPT Image 2, 1.5 and 1-mini generation and editing with your own API key: animated GIF, SVG and markup edits on 2.6, no credits middleman, exact USD cost per run
Custom ComfyUI tools and utilities. (Description by CC)
ComfyUI custom node that interfaces with Nano Banana Pro image generation API through Google AI Studio.
Custom ComfyUI output node that mirrors core save behavior with a boolean toggle for writing PNG metadata with prompt information.
This repository wraps the latest BiRefNet model as ComfyUI nodes. Compared to the previous model, the latest model offers higher and better matting accuracy.Support TensorRT.
Audio/video load and format-conversion nodes for ComfyUI. Reads and writes only inside the ComfyUI input/output directories; every path is resolved with realpath and confirmed with commonpath, so absolute paths, '..' traversal and outside symlinks are rejected. Extract the audio track from a video, trim long media by time range, export several formats in one run with parameters auto-adapted to the source, instant same-format passthrough with no re-encoding, a standalone video parameter report, and built-in compliance checks for mainstream Chinese music platforms. English UI by default; a Simplified Chinese translation is available from Interface -> Language. Files are read from the ComfyUI input directory (drag and drop, or the upload button on the load nodes) and written to the output directory under a relative prefix. Report text follows the interface language by default and can be pinned per node.
Lightweight ComfyUI node pack for storing and retrieving any data by name across workflows. (Description by CC)
Custom ComfyUI node for compositing material textures onto floors with perspective using SAM2 floor masks and perspective warping. (Description by CC)
A powerful prompt generation node designed for vision models that analyzes images and transforms visual features into high-quality prompts. (Description by CC)
Quality-of-life nodes for ComfyUI: Compare videos, character chat and cards, MiniMax H3 RefMod, optimised GIFs, run timer, paste stack, and plenty more
ComfyUI wrapper for Speech-to-Speech translation, hibiki: https://github.com/kyutai-labs/hibiki
Eigen AI FLUX API integration for ComfyUI with LoRA support and large font prompt inputs. Features high-quality image generation using FLUX.1-schnell model with multi-LoRA support, content upscaling, and optimized prompt input interface.
Architectural visualization upscaler with collapsible UI, dual ControlNet, and perspective correction
Backend-agnostic two-pass scene directing and H3 Ref2VA prompt composition for ComfyUI, with optional OBVPM timeline integration
A custom node that fetches Maplestory character images using Nexon API with configurable actions, emotions, and weapon motions. (Description by CC)
ComfyUI custom node for local translation with Gemma 4 E2B
ComfyUI custom node for local GGUF-only translation with Gemma 4
ComfyUI plugin providing auto-sized image and latent nodes with model-aligned resolutions.
Pick one item out of a ComfyUI list by index, with an index that can step once per queued job.
High-quality image dithering node pack supporting Atkinson and Halftone methods, with channel-split capabilities for RGB/CMYK printing effects.
Auto-shutdown timer, FileBrowser launcher, and Pod Info utilities for ComfyUI running on RunPod.
Local Ollama prompt enhancer nodes for video and audio direction, designed for LTX 2.3 talking-head workflows.
Custom nodes for ComfyUI that integrate with fal.ai's FLUX 2 and FLUX 1 LoRA APIs for text-to-image generation.
Nuke-style node graph for ComfyUI: Nuke hotkeys (Tab, F, D, Shift+Q, Ctrl+Shift+X…), straight Nuke pipes, groups as backdrops, remappable keys and presets. No Python nodes, zero dependencies.
ZX Spectrum pixel-art converter for ComfyUI. Based on spectrumify-public by JGF.
ComfyUI prompt-generation nodes backed by external LLM providers.
ComfyUI Storyboard Editor: grid preview, add/delete/edit, insert/replace, auto-expand panels. ComfyUI 分镜编辑器:宫格预览、增删改、插入替换、自动扩充宫格。
JosiaNodes 实验包:与主节点包分离的独立试验田,节点可随意增删。含 Qwen-Image 2.1 加速 LoRA 适配器(JosiaQwenAccel,Fun PDD / Viggle Turbo 两条路线)与 Josia框选编辑节点(图像上画框标注、导出官方格式 JSON)。
ComfyUI custom nodes for true multi-prompt batch conditioning.
A comprehensive ComfyUI plugin that integrates Google AI Studio's powerful AI models for content generation, image creation, video planning, and advanced prompt optimization.
Create a mask to slice the image at specific coordinates
Four-way alignment, travelling wave, and block-glitch video compositor nodes for ComfyUI.
ComfyUI open nodes by juntaosun.
Custom nodes to make ComfyUI more accessible, providing centralized control to manage bypass state of multiple nodes with organized groups.
A ComfyUI node back that makes it easy to save metadata for use with sites like CIVITAI.COM using two simple nodes.
Provides nodes to support generation of all possible combinations of a set of prompts via boolean logic
AI-written node to extract basic settings in workflow, with limitations on subgraphed nodes and math/logic node compatibility.
A single node that downloads a checkpoint, converts its DiT to Apple MLX, runs the semantic-first dual-time turbo denoise in MLX, and outputs an image — on Apple Silicon.
Simple Gemini API node & AR/Resolution calculator & Compositing
Download large ComfyUI models faster by fetching file pieces in parallel with HTTP range requests.
Load and use your Draggen.io moodboards directly in ComfyUI.