<h1 id="promptwaffle-wf-krea2-turbo-uncensored-production-pipeline-v3">Promptwaffle WF - Krea2 Turbo Uncensored Production Pipeline v3</h1><p>A streamlined ComfyUI workflow for <strong>Krea 2 Turbo</strong> with LoRA stacking, automatic trigger words, resolution presets, reproducible seeds, and embedded generation metadata.</p><h2 id="choose-your-checkpoint">Choose Your Checkpoint</h2><h3 id="uncensored-version">Uncensored Version</h3><p>For unrestricted local generation, use:</p><p><a target="_blank" rel="ugc" href="https://civarchive.com/models/2777907/my-krea2-unlocked-int8-convrot-diffusion-model?modelVersionId=3128223">My Krea2 Unlocked INT8 ConvRot</a></p><p>The workflow is configured for:</p><p><code>myKrea2UnlockedInt8_v10.safetensors</code></p><p>Place it in:</p><p><code>ComfyUI/models/diffusion_models/</code></p><h3 id="official-krea2-version">Official Krea2 Version</h3><p>Those who do not need the unlocked checkpoint can use the official ComfyUI Krea2 Turbo model:</p><p><a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/Krea-2/blob/main/diffusion_models/krea2_turbo_fp8_scaled.safetensors">Official Krea2 Turbo FP8 Scaled</a></p><p>Select the downloaded checkpoint in the <strong>Krea2 Diffusion Model</strong> loader.</p><h2 id="required-components">Required Components</h2><h3 id="text-encoder">Text Encoder</h3><p><a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/Krea-2/tree/main/text_encoders">qwen3vl_4b_fp8_scaled.safetensors</a></p><p>Place in:</p><p><code>ComfyUI/models/text_encoders/</code></p><h3 id="vae">VAE</h3><p><a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/Krea-2/tree/main/vae">qwen_image_vae.safetensors</a></p><p>Place in:</p><p><code>ComfyUI/models/vae/</code></p><p>The CLIP Loader type must be set to:</p><p><code>krea2</code></p><h2 id="required-custom-nodes">Required Custom Nodes</h2><h3 id="comfyui-lora-manager">ComfyUI-Lora-Manager</h3><p><a target="_blank" rel="ugc" href="https://github.com/willmiao/ComfyUI-Lora-Manager">https://github.com/willmiao/ComfyUI-Lora-Manager</a></p><p>Provides LoRA stacking, trigger-word controls, prompt encoding, metadata saving, and debugging.</p><h3 id="comfyui-resolution-master">Comfyui-Resolution-Master</h3><p><a target="_blank" rel="ugc" href="https://github.com/Azornes/Comfyui-Resolution-Master">https://github.com/Azornes/Comfyui-Resolution-Master</a></p><p>Provides resolution and aspect-ratio presets.</p><h2 id="workflow-features">Workflow Features</h2><ul><li><p>Separate style and character LoRA stacks</p></li><li><p>Automatic LoRA trigger-word injection</p></li><li><p>Trigger-word toggles</p></li><li><p>Optional inline LoRA loading</p></li><li><p>Resolution Master presets</p></li><li><p>Shared seed for image and wildcard reproduction</p></li><li><p>Final resolved-prompt preview</p></li><li><p>Prompt, seed, model, LoRA, and workflow metadata embedded in saved PNGs</p></li><li><p>Drag generated PNGs back into ComfyUI to restore the workflow</p></li></ul><h2 id="recommended-settings">Recommended Settings</h2><ul><li><p><strong>Steps:</strong> 8</p></li><li><p><strong>CFG:</strong> 1.0</p></li><li><p><strong>Sampler:</strong> Euler</p></li><li><p><strong>Scheduler:</strong> Simple</p></li><li><p><strong>Denoise:</strong> 1.0</p></li></ul><p>Krea2 Turbo is distilled, so increasing CFG above 1.0 usually makes results worse.</p><p>There is intentionally no traditional negative prompt. The workflow uses <strong>Conditioning Zero Out</strong>, which is appropriate for CFG 1.0.</p><h2 id="prompting">Prompting</h2><p>Edit the node labeled:</p><p><strong>User Prompt — EDIT THIS</strong></p><p>Krea2 works best with natural-language prompts describing the subject, action, environment, lighting, camera angle, mood, and important details.</p><p>Avoid older tag-heavy prompting styles such as <code>masterpiece</code>, score tags, or excessive weighting syntax.</p><h2 id="lora-strengths">LoRA Strengths</h2><p>Useful starting values:</p><ul><li><p>Style LoRA: <strong>0.75–0.9</strong></p></li><li><p>Character LoRA: <strong>0.6–0.8</strong></p></li><li><p>Secondary stacked LoRA: lower it if the LoRAs begin competing</p></li></ul><p>The included LoRAs are examples and can be replaced with your own Krea2-compatible LoRAs.</p><h2 id="resolution">Resolution</h2><p>Good starting resolutions include:</p><ul><li><p>1024 × 1024</p></li><li><p>1216 × 832</p></li><li><p>832 × 1216</p></li><li><p>1328 × 1328</p></li></ul><p>Generate near the model’s native resolution and upscale afterward for larger final images.</p><p>Built by <strong>Promptwaffle</strong> for fast Krea2 LoRA testing, repeatable generations, and everyday production use.</p>

Files with SHA256: 50a40bdbbb07039f29355183d6cf63037dcbaf1f9c40dc75c455b56171c464e1