ComfyUI, a company focused on empowering creators to manage image, video, and audio outputs via diffusion models with a node-based system, has successfully secured a $30 million investment round, bringing its valuation to $500 million.

The investment round was spearheaded by Craft Ventures, with contributions from various other backers such as Pace Capital, Chemistry, and TruArrow.

Founded in 2023 as an open-source initiative, ComfyUI emerged shortly after the debut of diffusion models. At that point, prototypes like Midjourney and OpenAI’s DALL-E were unreliable, often making glaring errors, for example, incorporating extra fingers on hands.

To overcome these issues, the founders crafted a modular system that offers creators detailed control over every phase of the generation process.

Their solution attracted substantial attention among creative professionals, culminating in the formation of a formal startup. In late 2024, ComfyUI secured $19 million in Series A funding from entities including Chemistry Ventures, Cursor Capital, and Guillermo Rauch, founder of Vercel.

While the latest diffusion models have significantly improved from adding a sixth digit to hands, the demand for the kind of precise control that ComfyUI provides has only intensified.

“When you consider your average prompt-based solution, like Midjourney or ChatGPT, you typically achieve around 60% – 80% of your goal,” Yoland Yan, co-founder and CEO of ComfyUI, explained to TechCrunch. “However, to adjust that final 20%, you have to keep trying various options.”

Yan (shown left) equated the experience to gambling, as prompting the model for a minor adjustment could yield an entirely different result, potentially overwriting previously satisfactory elements.

ComfyUI’s node-based interface enables creators to connect specific elements of the generation workflow, providing them with comprehensive control over the quality of their outcomes.

“That message cannot be easily captured in the prompt box [of a foundational model],” Yan noted.

Feedback from users indicates agreement, as ComfyUI reports having over 4 million active users.

Creative professionals are utilizing the tool for applications in visual effects, animation, advertising, and even industrial design.

The startup asserts that its solution has become an indispensable asset for technical artists and other creatives, making it common to see “ComfyUI artist or engineer” as a job title on studio employment listings.

Even with ongoing advancements in video and image foundational models, Yan argues that they are still not ideal, ensuring that a tool like ComfyUI will remain in high demand.

“In a landscape where AI mistakes are going to proliferate, the Comfy approach of incorporating human input will ultimately capture the majority of interest,” he stated.

Competitors to ComfyUI include Weavy, a startup recently acquired by Figma.