krea2

AI filmmaking on a node canvas. Generate locally on your own GPU with the Inline Core engine (Z-Image, Krea 2, FLUX.2, MiniMax H3), or use hosted models with no GPU. Train your own LoRAs locally on the same canvas. Every render is kept as a versioned take, so you never lose the good version. Free and open source.

Krea 2 & Klein 9B LoRA - LoKR Studio — train, profile, repair, and extract Krea 2 & Flux 2 Klein 9B LoRAs & LoKRs

Advanced Seed Diversity, Conditioning Enhancement and Step Control node for ComfyUI

Native ComfyUI nodes for Krea 2 AnyPaint - arbitrary-mask inpainting, outpainting, and mixed image editing.

475 tested Krea 2 Turbo prompts. One file, drop it in ComfyUI.

Prompt-only Krea moodboard search and style nodes for ComfyUI.

StormForge Krea2 Turbo LoRA Merger Suite

ComfyUI custom nodes that guide Krea 2 with reference images: stack up to 12 guide cards using plain-language recipes - keep the subject, borrow style/palette/lighting/material, copy pose or shapes, avoid text/logos.

Run Krea 2 (Krea AI's 12B DiT) natively in Forge Neo — the first open-source Krea 2 extension. One-click model download, presets, fp8, full + piecewise loading. By Stable Yogi.

Prompt builder for Krea-2 in ComfyUI. Fill structured fields and get flowing prose in the format the model expects, not a keyword list.

Lightweight LOWVRAM optimised KREA2 webui

A ComfyUI diffusers implementation of Krea2 for the highest possible quality and most options

Regional prompting + per-region LoRA for Krea 2 in ComfyUI — single-pass, canvas editor

Z-Image, QWEN, Flux, ERNIE, Krea-2 , inference with ComfyUI

Tooling for dataset captioning, model merging and quantizing.

Identity-preserving image editing for Krea 2 in Diffusers

musubi-tuner GUI (Krea2 only)

The full Krea 2 Enhancement Suite for Forge Neo — advanced Prompt-Adherence engine + custom per-layer Detail Boost PRO. Free companion to sd-forge-krea2. By Stable Yogi.

Single-user LoRA training and generation platform on Modal. Train Krea 2 LoRAs, generate stills, and turn them into video with MiniMax-H3 (sound and picture in one pass) or Wan 2.2 — one deploy, one URL, no infrastructure to keep alive.

Model-aware sigma schedules, sampler compatibility validation, and versioned inference profiles for ComfyUI