ComfyUI Custom Extensions

Local Fish Audio S2 Pro GGUF speech generation in ComfyUI via s2.cpp, with multi-speaker and voice-management nodes.

A powerful tool for AI dynamic comic creators and scriptwriters extracting up to 4 specific keys from JSON scripts.

ComfyUI custom node for DINOv3 point prompt functionality with integration of Facebook's DINOv3 research model. (Description by CC)

Lightweight ComfyUI custom nodes for working with image and mask batches. No dependencies beyond ComfyUI itself.

ComfyUI custom nodes for merging multiple GLB files and interactively controlling sub-models within a single GLB.

X-WIDE Prompt Input Helper —— 正向/反向提示词浮窗：按键（默认 P / N，可自设）弹出输入框，从别处复制的内容自动收进来，回车直接写回画布上那个文本控件，不新增节点、不改现有流程

Unofficial implementation of 'Classifier-Free Guidance: From High-Dimensional Analysis to Generalized Guidance Forms' for ComfyUI. (cui-ggf for short)

Telegram bot for driving ComfyUI: txt2img/img2img, LoRA, ControlNet, workflow parameters — from chat

ComfyUI nodes for API-based inpainting (Gemini, Imagen) with aspect ratio constraints, smart cropping, resize fitting, intelligent paste-back with transparency support. (Description by CC)

Save ComfyUI images to S3-compatible storage with local previews, public links, and a native sidebar.

Load Image that lists sub-directories of ComfyUI/input, with a folder dropdown, a root-only filter, working sub-directory previews, cyclic arrow navigation, and a lightweight built-in mask editor.

Unopinionated llama.cpp bindings for ComfyUI with generic parameters and multi-stage chat support for LLM integration.

Custom node pack for embedding draft and production modes within ComfyUI workflows using if-statement-like conditional branching. (Description by CC)

Custom node providing text preservation preprocessing for ComfyUI using OCR to identify text regions and apply blur/noise fusion to non-text areas. (Description by CC)

ComfyUI custom nodes for one-sample-per-run iteration across image, video, and text sources.

Simple HQ file save for audio workflows supporting studio quality 24-bit 192kHz or higher on compute devices with optional DSP support for optimization/offload.

Experimental custom node for audio VST plugins using pedalboard for audio composition generators (AudioLDM, AceStudio XL, etc.)

A custom node for ComfyUI designed to simplify the 'High-Resolution Fix' process. This node takes an input image, performs a direct upscale, and then re-encodes it into a latent image, preparing it for a final sampling pass to add details.

A ComfyUI plugin providing agent integration nodes for base64 image processing without temporary files, supporting PNG, JPEG, and WEBP formats with adjustable compression quality. (Description by CC)

ComfyUI plugin providing vision-language model (VLM) conditioning via Qwen3-VL for Z-Image-Turbo video generation with multi-modal concept consistency.

Experimental per-model native Aiter attention for AMD ROCm. Requires a compatible native Aiter FMHA build; tested on Windows gfx1151.

ComfyUI custom node for PuLID v1.0/v1.1 with automatic version detection that eliminates user version awareness requirements. (Description by CC)

Collection of simple utility nodes for ComfyUI use, including prompt normalization, text concatenation, Korean image prompt generation, and image processing. (Description by CC)

Simplified Gemini 2.5 Flash Image Preview node for ComfyUI. REST-only for stability, two optional reference images, padded aspect ratio resizing (no stretching), and basic sampling controls. All extra debug layers, SDK path, multi-seed, and legacy compatibility code removed to avoid crashes.

Content-addressed image batch planning, verified saves, failure tracking and selective retry.

Chinese commercial-copy verification with OCR evidence, field matching, review verdicts and repair masks.

Inspect aligned image edits against allowed/protected regions and restore original protected pixels.

Locate heuristic video exposure/chroma flicker and near-identical frame runs with scene exclusions and frame-indexed reports.

Explicit IMAGE/MASK canvas, batch, numeric and coverage contracts with per-frame evidence and strict gating before inpaint or composite.

Strict prompt-template input contracts with bounded JSON validation and deterministic one-pass rendering.

Rank multiple continuation clips by visual seam continuity and recommend duplicate-frame trims.

Capture local workflow output baselines and inspect image, ROI, environment and timing regressions.

A collection of custom nodes for using the Stability AI API in ComfyUI.

Structure-guided three-tone color reduction, source color restoration, and production pixel-art cleanup nodes for ComfyUI.

ai-driven toolchain for digital design and fabrication

some comfyui custom nodes to make effect shatter

some comfyui custom nodes to set it as known type

comfyui custom nodes to caption image with joy

Save HDR-encoded IMAGE batches (e.g. NVIDIA RTX TrueHDR output) as 10-bit HEVC HDR MP4 (HLG or PQ/HDR10) with correct BT.2020 color tagging and HDR10 metadata.

ComfyUI plugin for Sa2VA segmentation with independent model loaders, ComfyUI memory management, and VITMatte edge refinement. (Description by CC)

ComfyUI custom node for toggling prompt groups by title from JSON files with flexible prompt selection and output formatting. (Description by CC)

Dynamic switch nodes for ComfyUI providing flexible iterative input selection with intelligent fallback logic.

MP4/MOV to a single IMAGE frame — loop, random, skip-every-nth, frame counters (ib video slicer).

Single-cell analysis nodes for ComfyUI.

Run takecheck's acceptance checks on a generated take: frozen frames, duplicate frames, blank frames, flicker, motion, subtitle residue and audio. CPU only.

A simple ComfyUI utility node that calculates image width and height from a target longest-side value and an aspect ratio.

A simple ComfyUI custom node that inspects workflow in PNG file and loads embedded metadata. (Description by CC)

ComfyUI node for LLM client integration. (Description by CC)

Images, video, speech, music and text from 450+ AI models (Veo, Kling, Seedance, Nano Banana, GPT Image, FLUX, Claude, GPT and more) with one GenMagic API key and one pay-as-you-go balance. Image-to-image and image-to-video included.

VRAM purge button in the ComfyUI menu - clears ComfyUI from VRAM without restarting

A custom node for ComfyUI that transforms user photos into stylized character avatars.

Personal custom nodes repository for ComfyUI including Mosaic By Mask and Censor Bars By Mask nodes. (Description by CC)

A ComfyUI frontend extension for hiding and rearranging left sidebar toolbar icons, with optional auto-hide.

Load images sequentially or in batch from a specified folder.

Custom ComfyUI node providing unified access to multiple Large Language Models including OpenAI-compatible APIs and local Ollama instances with image support.

ComfyUI custom node for Gemini image generation via OpenRouter API, supporting text prompts and multiple image inputs (up to 4 images) with automatic saving to output directory. (Description by CC)

Next to AI mathematical methods can be used for the detection of objects like a circle.

ComfyUI simple square masks creation.

Custom nodes for ComfyUI with specialized image processing and saving functionality.

ComfyUI node that builds a VintageAnima caption from series, cast, designer, scan, and quality.

ZZNodes tools for ComfyUI: Windows shutdown after 30 seconds of queue idle, with cancellation and on-node logs.

A smart logic switch for ComfyUI that physically mutes upstream nodes to save GPU/CPU resources, featuring dynamic inputs and auto-cleanup.

Memory-conscious ComfyUI nodes used by TUTU's LTX 2.3 360 VR video workflow.

NODES: Prompt Preset, Video Batch Loader, Video Combiner

Comfyui-SceneDetect is a ComfyUI custom node that uses PySceneDetect to locate scene boundaries in a video and emit one representative frame per scene as an `IMAGE` batch. It also returns per-scene metadata as JSON (`STRING`) and the total number of detected scenes (`INT`).

comfyui_browser_url_workflow is custom node that open workflow from browser url params.

A ComfyUI custom node that scales images proportionally and fills to specified dimensions with a solid color background, supporting size reduction and centered placement. (Description by CC)

use TAESD decoded image.you need donwload taesd_decoder.pth and taesdxl_decoder.pth to vae_approx folder first. It will result in a slight loss of image quality and a significant decrease in peak video memory during decoding.

Comprehensive image processing plugin for ComfyUI

comfyui个人插件

ComfyUI custom nodes for calling OpenAI's DALL-E 3 image model through a Zuco-compatible base URL with text-to-image and image-to-image support.

Gemini chat and image generation via LiteLLM. Multimodal support with multiple reference images. Precise control over image resolution (1K/2K/4K) and aspect ratio (1:1, 16:9, 4:3, 9:16, 3:4). Temperature control (0-1). Zero external dependencies.

Precise video object removal in ComfyUI, with masks, clean-frame guides and an optional local Studio interface.

Displays real-time GPU usage, VRAM consumption and temperature inside ComfyUI UI.

ComfyUI custom node for image quality evaluation — FID / IS / CLIP-Score / LPIPS

NODES: Item List, Template Input, Template Processor, Scene Queue Node, Trigger Passthrough Node