ComfySearch: Autonomous Exploration and Reasoning for ComfyUI Workflows

2026-01-07

arXiv.org

AI Summary

This work addresses the challenges in ComfyUI workflow construction—namely, low executability and suboptimal quality—stemming from its large component set and stringent graph-structure constraints. To overcome these limitations, we propose an agent-based, verification-guided exploration framework that integrates dynamic validation feedback with graph-constrained reasoning and autonomous search strategies. This approach enables the efficient generation of high-quality, executable, and modular workflows while preserving long-range structural consistency. Experimental results demonstrate that our method significantly outperforms existing approaches on complex creative tasks, achieving substantial improvements in workflow executability, task success rate, and generalization capability.

Technology Category

• Constraint Satisfaction and Optimization: Other Foundations of Constraint Satisfaction
• Cognitive Modeling & Cognitive Systems: Computational Creativity
• Knowledge Representation and Reasoning: Computational Complexity of Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Economics, Online Markets and Human Computation: Architectures and workflows that use LLMs for crowd work
• Search and Retrieval-Augmented AI: Agentic search

Abstract

AI-generated content has progressed from monolithic models to modular workflows, especially on platforms like ComfyUI, allowing users to customize complex creative pipelines. However, the large number of components in ComfyUI and the difficulty of maintaining long-horizon structural consistency under strict graph constraints frequently lead to low pass rates and workflows of limited quality. To tackle these limitations, we present ComfySearch, an agentic framework that can effectively explore the component space and generate functional ComfyUI pipelines via validation-guided workflow construction. Experiments demonstrate that ComfySearch substantially outperforms existing methods on complex and creative tasks, achieving higher executability (pass) rates, higher solution rates, and stronger generalization.

Problem

• ComfyUI
• workflow generation
• structural consistency
• component space
• executability

Innovation

• ComfySearch
• autonomous exploration
• validation-guided workflow construction
• ComfyUI workflows
• agentic framework