All News DISPATCH WORKFLOW

Krea 2 Image model integrates with ComfyUI via Partner Nodes

The Krea 2 Image model is now available as a native node within the ComfyUI ecosystem. This integration allows creators to combine Krea’s aesthetic control with the granular automation of node-based workflows.

ComfyUI

ComfyUI, the node-based interface for Stable Diffusion and generative media, has integrated the Krea 2 Image model through its new Partner Nodes system. This update allows artists to run Krea’s proprietary image generation model directly within their custom workflows rather than using the Krea web interface. The integration focuses on providing creators with specific control over the aesthetic direction and visual consistency of generated frames.

What's new

The Krea 2 Image model is accessible via a dedicated node in ComfyUI as of February 2025. This implementation uses the Krea API, meaning the heavy lifting of the generation happens on Krea’s servers while the logic and pipeline remain inside the user's local or cloud-based ComfyUI environment. The node supports standard parameters including prompt input, aspect ratio selection, and seed control, but it is specifically tuned for the 'look and feel' adjustments that Krea is known for in the creative community.

By using the Partner Node, ComfyUI users can now pipe the output of a Krea 2 generation directly into other nodes for upscaling, face restoration, or ControlNet passes. This eliminates the need to manually download images from Krea and re-upload them into a local workflow, reducing friction for professional production pipelines.

How it fits your workflow

ComfyUI users often struggle with the 'raw' look of base models like SDXL or Flux.1, which frequently require complex LoRA stacks to achieve a specific cinematic or commercial aesthetic. The Krea 2 Image model serves as a pre-tuned alternative that prioritizes high-end visual output without the need for extensive prompt engineering. For concept artists and mood-board creators, this provides a faster route to a polished image compared to standard Stable Diffusion setups.

In a production environment, this integration makes ComfyUI a viable competitor to all-in-one platforms like Leonardo.ai or Midjourney. While Midjourney offers high aesthetic quality, it lacks the workflow automation of ComfyUI; conversely, while ComfyUI offers total control, it can sometimes lack the 'out-of-the-box' beauty of proprietary models. Krea 2 Image bridges this gap. Editors and VFX artists can use Krea 2 for initial asset generation and then use ComfyUI’s native tools for precise image-to-image tasks or batch processing.

What it costs / how to try it

Accessing the Krea 2 Image node within ComfyUI requires a Krea account and an active API key. Users can install the node through the ComfyUI Manager by searching for 'Krea' under the Partner Nodes section. Usage is billed according to Krea’s current API pricing tiers, which are separate from any local hardware costs.

Read the original announcement on ComfyUI ↗

Powered by ReelStack

Help keep this running

Your tip funds servers, models, and the time it takes to ship new tools faster. Set any amount below — every bit helps.