A fast and flexible ComfyUI workflow for Krea 2 Turbo designed for high-quality photorealistic image generation and easy LoRA integration.The workflow is built to keep the setup simple while giving you direct control over the most important generation parameters.<h3 id="features">✨ Features</h3><ul><li>⚡ Krea 2 Turbo generation</li><li>🎨 LoRA support with adjustable strength</li><li>📝 Simple prompt input</li><li>🤖 Optional AI prompt enhancement</li><li>🎯 Customizable seed</li><li>📐 Built-in aspect ratio and resolution selector</li><li>🔧 Adjustable LoRA trigger word</li><li>🔄 Easy switching between base model and LoRA</li><li>🖼️ Automatic image scaling</li><li>💾 Automatic image saving</li><li>🚀 8 steps / Euler / Simple scheduler default configuration</li><li>✨ Krea 2 conditioning rebalance for improved prompt control</li></ul><h3 id="prompt-enhancement">🧠 Prompt Enhancement</h3>The workflow includes an optional AI-powered prompt enhancement system.You can simply enter a short description of what you want to generate and let the built-in text generation node expand it into a more detailed image-generation prompt.Prompt enhancement can be enabled or disabled with a switch.Prefer writing your own prompts?Just turn it off and use your prompt directly.<h3 id="lora-support">🎨 LoRA Support</h3>The workflow includes a dedicated LoRA system with:<ul><li>LoRA enable/disable switch</li><li>LoRA selection</li><li>Adjustable LoRA strength</li><li>Custom trigger word</li><li>Easy switching between the base Krea 2 model and LoRA</li></ul>This makes the workflow especially convenient for working with character LoRAs.<h3 id="resolution">📐 Resolution</h3>A built-in resolution selector makes it easy to change:<ul><li>Aspect ratio</li><li>Megapixel count</li><li>Output dimensions</li></ul>The default workflow is configured for 1:1 at 4 megapixels, but you can adjust the resolution to your needs.<h3 id="default-settings">⚙️ Default Settings</h3>Model: Krea 2 Turbo Text Encoder: Qwen3 VL 4B FP8 VAE: Qwen Image VAE Steps: 8 CFG: 1 Sampler: Euler Scheduler: Simple Denoise: 1.0The included workflow uses Krea 2 Turbo INT8 ConvRot for efficient generation.<h3 id="required-models">📦 Required Models</h3>You will need the following Krea 2 components:<ul><li><code>krea2_turbo_int8_convrot.safetensors</code></li><li><code>qwen3vl_4b_fp8_scaled.safetensors</code></li><li><code>qwen_image_vae.safetensors</code></li></ul>Place them in the appropriate ComfyUI model directories.<h3 id="example">💜 Example</h3>The workflow is configured with Amara Sterling Krea 2 LoRA as an example, but you can replace it with any compatible Krea 2 LoRA.The included example demonstrates how to generate a photorealistic character with a short natural-language prompt and the LoRA system.<h3 id="free-ai-generator">🤖 Free AI Generator</h3>Want to try AI image generation without setting up ComfyUI?🎁 Free generations: @drgirls_ai_bot<a target="_blank" rel="ugc" href="https://t.me/drgirls_ai_bot">https://t.me/drgirls_ai_bot</a><h3 id="more-from-novaailabs">💜 More from NovaAILabs</h3>Creator: <a target="_blank" rel="ugc" href="https://civarchive.com/user/NovaAILabs">https://civarchive.com/user/NovaAILabs</a>NovaAILabs Collection: <a target="_blank" rel="ugc" href="https://civarchive.com/collections/17587706">https://civarchive.com/collections/17587706</a>I regularly create photorealistic characters, LoRAs, workflows and prompt packs for AI image generation.If you find this workflow useful, feel free to leave a rating and share your creations.

Files with SHA256: d43fb91dcb190e16912a8d4f0ebff32291cdb69da5a0df42a39e0f712ab9a8cc