ComfyUI reaches $500M valuation amid rising demand for creator-controlled AI tools

ComfyUI, a startup focused on helping creators manage image, video, and audio outputs from diffusion models through a node-based workflow, has secured $30 million in new funding, bringing its valuation to $500 million.

The funding round was led by Craft Ventures, with participation from investors including Pace Capital, Chemistry, and TruArrow.

ComfyUI began as an open-source project in 2023, emerging shortly after the rise of diffusion-based AI models. During that period, systems such as Midjourney and DALL-E were still in their early stages. They often produced flawed outputs, including visual errors such as extra fingers in generated images.

To overcome these issues, the founders built a modular framework that gives creators precise control over each stage of the generation process. This approach allowed users to refine outputs more effectively than traditional prompt-based systems.

The project quickly gained traction among creative professionals and eventually transitioned into a full-fledged startup. In late 2024, ComfyUI raised $19 million in a Series A round backed by investors including Chemistry Ventures, Cursor Capital, and Guillermo Rauch.

Despite significant improvements in modern diffusion models, the need for detailed control has continued to grow. According to co-founder and CEO Yoland Yan, prompt-based tools often fall short of delivering complete results. He explained that such systems typically achieve only about 60% to 80% of the desired outcome, leaving users to experiment repeatedly to refine the remaining details.

Yan described this process as similar to a slot machine, in which adjusting one element can unintentionally alter parts of the output that were already correct.

ComfyUI addresses this challenge through its node-based interface, enabling users to connect and control specific components of the generation workflow. This structure yields more consistent and predictable results than relying solely on prompts.

“You cannot easily convey that message in the prompt box of a foundational model,” Yan noted.

The platform’s adoption reflects this demand, with ComfyUI reporting more than 4 million users.

Its tools are now widely used by professionals across industries, including visual effects, animation, advertising, and industrial design. The company notes that the platform has become so integral to certain workflows that job listings increasingly include roles such as “ComfyUI artist” or “ComfyUI engineer.”

Although foundational models for image and video generation continue to advance, Yan maintains that they are not yet perfect and that tools offering fine-grained control will remain essential.

“In the world where AI-generated content is everywhere, a human-in-the-loop approach like ComfyUI’s will ultimately capture the most attention,” he said.

ComfyUI operates in a competitive space that includes companies such as Weavy, which Figma acquired last year.