AP Workflows for ComfyUI (APW) - All-in-One Media Generation Solution

What is AP Workflows (APW) for ComfyUI?

APW is a comprehensive, enterprise-grade, all-in-one solution that enables you to build your own private, Midjourney-like image and video generation system. Created by Alessandro Perilli, APW serves as a blueprint for automated AI media production, providing carefully curated nodes, pre-set models, and optimized workflows that transform ComfyUI into a powerful production system without requiring technical expertise.

Professional and Industrial Use Cases for APW

AP Workflows are used by professional studios and industrial-scale organizations worldwide for production-ready applications:

• Creative Design: Fast idea prototyping for interior, industrial, and game design
• Film Production: AI Cinema image generation for storyboarding and location scouting
• Fashion Industry: Creative upscaling of images and remote model casting for collections and advertisements
• Visual Effects: Face swap for stunt scenes and vintage film restoration
• E-commerce: Virtual try-on for online shopping and retail point of sale
• Art Generation: Complex image creation with consistent styles and themes

Whether you're a fashion brand, e-commerce company, photography studio, ad agency, design studio, gaming company, AI-first startup, or innovative artist, APW helps incorporate AI into your professional creative workflow at industrial scale—turning experimental AI tools into production-ready systems.

Benefits of APW

• Complete All-in-One Solution: A carefully designed workflow with curated sets of the best nodes from the AI community
• Industrial Scale Production: Generate and manipulate media at scale for commercial applications
• Unified Generation System: Integrated pipelines for images (FLUX, SD3, SD1.5/SDXL) and videos (Hunyuan, CogVideoX)
• Advanced Enhancement Tools: Face and hand detailing, object swapping, face swapping, and high-quality upscaling
• Prompt Intelligence: LLM and VLM integration to improve prompts and generate sophisticated captions
• Ready-to-Use Cloud Environment: Easily run the workflow in a cloud environment without worrying about setup or installation (for local installation, please refer to the )

AP Workflow Overview

APW is a large, sophisticated workflow designed to give digital artists, filmmakers, and advertisers unprecedented control over AI generation. While its size may initially seem daunting, the workflow is organized into logical functional groups with a centralized Controller system.

Main Components:

1. Central Controller: Single hub to enable/disable all workflow functions
2. Prompt Enhancement: LLM-powered tools to generate better prompts and captions
3. Unified Generation Pipelines: Integrated systems for both image (FLUX, SD3, SD1.5/SDXL) and video content (Hunyuan, CogVideoX)
4. Comprehensive Manipulation: Tools for enhancing both images and videos, including face detailing, hand fixing, and object swapping
5. Creative Upscaling: Professional-grade upscalers comparable to commercial solutions
6. Image Finishing: Color correction, LUT application, grain effects, and watermarking
7. Evaluation Tools: Compare results and score media on aesthetic quality

The workflow is comprehensively documented and features an intelligent navigation system with bookmarks for easily moving between sections. For navigation shortcuts and other interface details, see the .

Getting Started with APW

APW is pre-configured to generate images with the FLUX 1 model. After opening the workflow, you can immediately queue a generation with the prompt already set in the Prompt Builder function to have a quick test.

Check the outcome of your image generation/manipulation in the Final Image Saver section at the bottom-left of the workflow.

Key Components and Functions

Central Controller

A unified control hub that allows you to enable/disable all workflow functions from a single location. This centralized approach makes managing the complex workflow simple and intuitive.

Prompt Enhancement System

Prompt Builder

A visual and flexible prompt builder that allows quick switching between frequently used types and styles for positive prompts, and frequently used negative prompts.

Prompt Enricher

Enriches your positive prompt with additional text generated by a large language model (LLM), supporting both centralized proprietary models and open access models installed locally.

Caption Generator

Automatically captions uploaded images using visual language models like Florence 2, OpenAI GPT-4V, or locally installed VLMs.

Unified Media Generation System

APW's integrated generation system handles both images and videos within a cohesive framework:

Comprehensive Generation Capabilities

• Image Generation: Three pipelines (FLUX.1, SD3, SD1.5/SDXL) that can work individually or in combination
• Video Creation: Professional-quality video generation with Hunyuan (T2V/V2V up to 1280x720px/129 frames) and CogVideoX (T2V/I2V up to 1360x768px/81 frames)
• Cross-Modal Generation: Seamless workflows between image and video, allowing for image-to-video and video-to-video transformations

Enhancement Tools

Image and Video Painters

• Repainter: Full image regeneration with controllable fidelity
• Inpainter: Targeted editing of specific image areas
• Image Expander: Extend image boundaries in any direction

Comprehensive Manipulation

• Object Swapper: Identify and replace specific objects within images
• Face Swapper: Replace faces while maintaining expression and lighting
• Face Detailer: Enhance facial features with separate treatment for small and large faces
• Hand Detailer: Fix and improve hand anatomy in generated images

Creative Upscaling

• CCSR Upscaler: SD 2.1-based upscaler for intricate pattern preservation
• SUPIR Upscaler: SDXL-based upscaler with creative options and LoRA support

Image Finishing

Final enhancement tools including colorization, color correction, LUT application, film grain effects, and watermarking.

Advanced Conditioning

LoRAs

Apply style and content modifications through LoRAs for all supported models, organized by categories for easier selection.

ControlNet

Use up to four concurrent ControlNet models for precise control over generations, with support for various preprocessors.

IPAdapter & Redux

Generate variants and compositions from reference images, allowing for consistent character generation and style transfer.

Evaluation Tools

• Image Comparer: Compare multiple image versions side-by-side
• Face Analyzer: Automatically select images with facial landmarks similar to references
• Aesthetic Score Predictor: Rank images based on aesthetic quality
• Debugging Tools: Extensive logging and preview functions for workflow understanding

Latest Features

APW is continuously updated with new capabilities. For details on the latest features in APW 12.0, including video generation enhancements, SD3.5 support, and workflow improvements, see the .

Credits and License

AP Workflows are designed by Alessandro Perilli. For more information, updates, and installation instructions, visit .

Special thanks to all the custom node creators mentioned in the workflow documentation, including: @rgthree @kijai @Kosinkadink @matan1905 @cubiq @ltdrdata @glibsonoran @crystool @huchenlei