ComfyUI-AngelosKar

<p align="center"> <img src="icon.png" width="160" height="160" alt="ComfyUI-AngelosKar Logo"> </p>

ComfyUI-AngelosKar

A clean, modular custom node pack for ComfyUI focused on automated batch image processing, modern folder selection, multi-garment VTON workflow control, and dynamic canvas tag toggling.

---

🌟 Features

• Sequential Image Scheduling: Process entire folders chunk-by-chunk using native ComfyUI Run on Change without moving or modifying files.
• VRAM-Safe List Execution: Outputs images as Python lists rather than heavy 4D tensors, enabling ComfyUI list mapping to process images individually.
• Canvas Tag Controllers: Dynamically Bypass or Mute canvas nodes simply by adding tags (e.g., [RMBG], [PREVIEW]) to their titles.
• Multi-Garment VTON Mode Control: Switch between 1 to 5 garments dynamically with automatic canvas muting for [G2], [G3], [G4], [G5].
• Native Windows Folder Dialog: Open modern Windows Explorer folder picker directly from the node.
• One-Click Counter Resets: Interactive frontend buttons to instantly reset batch progress without restarting ComfyUI.

---

📦 Nodes & Categories

1. AngelosKar/Batch

• 📊 AK Image List Scheduler (AK_ImageListScheduler)

  • Schedules slice-by-slice image runs across queue executions.
  • Inputs: folder_path, list_size, file_pattern, sorting (natural or alphabetical), load_remainder (on/off), remainder_position (start/end).
  • Has an interactive 📂 Browse Folder button and 🔄 Reset Counter button.
  • Outputs: folder_path, sorting, file_pattern, start_index, list_size, total_images, run_counter, total_runs, folder_analysis.
  • Designed to be connected directly to 📦 AK Image List Loader.

• 📦 AK Image List Loader (Scheduler) (AK_ImageListLoader)

  • Worker node that receives inputs from the Scheduler and loads the scheduled slice of images.
  • Outputs images as an IMAGE list (OUTPUT_IS_LIST = True), allowing downstream nodes to execute one-by-one with full resolution preservation.

• 🔄 AK Auto Image List Loader (AK_AutoImageListLoader)

  • All-in-one standalone node combining both the Scheduler and Loader into a single node.
  • Ideal for simple, compact batch workflows.

• 🔄 AK Reset Counters (AK_ResetCounters)

  • Utility canvas button. Place it anywhere on the canvas and click 🔄 Reset All Counters to reset all scheduler counters in the active workflow back to 0.

---

2. AngelosKar/Control

• 🔀 AK Tag Bypasser (Active / Bypass) (AK_TagBypasser)

  • Switches all canvas nodes matching a specified tag (e.g. [RMBG], [UPSCALE], [FACE]) between ACTIVE and BYPASS (passthrough).

• 🔇 AK Tag Muter (Active / Mute) (AK_TagMuter)

  • Switches all canvas nodes matching a specified tag (e.g. [PREVIEW], [SAVE], [EXTRA]) between ACTIVE and MUTE (disabled / skipped).

Canvas Priority Rule: If multiple controllers target the same node, the strict hierarchy is MUTE > BYPASS > ACTIVE.

---

3. AngelosKar/VTON

• 👔 AK Garment Mode Controller (1-5) (AK_GarmentModeController)
  • Designed for multi-garment Virtual Try-On (VTON) workflows.
  • Select active garments from 1 Garment up to 5 Garments.
  • Automatically mutes nodes tagged with [G2], [G3], [G4], or [G5] when that garment number is disabled. Nodes without garment tags are left untouched.

---

4. AngelosKar/Utils

• 📂 AK Folder Path Picker (AK_FolderPicker)
  • Interactive node with a 📂 Browse Folder button that opens the native Windows Explorer folder selection dialog and outputs the selected folder path string.

---

🚀 Installation

Option 1: ComfyUI-Manager (Recommended)

1. In ComfyUI, open ComfyUI-Manager.
2. Search for ComfyUI-AngelosKar.
3. Click Install and restart ComfyUI.

Option 2: Comfy-CLI

comfy node install comfyui-angeloskar

Option 3: Manual Git Clone

1. Open a terminal in your ComfyUI/custom_nodes/ directory.
2. Clone the repository:

   git clone https://github.com/AngelosKar-code/ComfyUI-AngelosKar.git
   

3. Restart ComfyUI.

---

🛠 Requirements

• Python >= 3.10
• ComfyUI (frontend v1.x or v2.x)
• Dependencies:
  • torch
  • numpy
  • Pillow
  • aiohttp

---

📄 License

This project is licensed under the MIT License.