<h1 id="flux-kontext-character-expansion-workflow">🧩 Flux-Kontext Character Expansion Workflow</h1><h3 id="overview">Overview</h3>This custom Flux-Kontext workflow is a major extension of the base model, built to handle multi-image composition, LoRA stacking, and multi-step generation for high-quality, consistent character datasets. It’s designed for character creation, LoRA training, and regeneration of low-quality inputs — giving you complete control over appearance, context, and detail. The NSFW version is here: <a target="_blank" rel="ugc" href="https://civarchive.com/models/2088506">https://civarchive.com/models/2088506</a> It uses line art refernces with inpainting of the subject - so you find any position you like and your reference image will be painted into it. Do you want ALL the NSFW LoRAs, checkpoints, and Workflows? Here is my patreon - i will uploading some CRAZY stuff I have been working on. <a target="_blank" rel="ugc" href="https://www.loboforge.com/@loboforge">https://www.loboforge.com/@LoboForge</a><hr /><h3 id="dual-image-context-system">🖼️ Dual-Image Context System</h3>The workflow allows you to combine two different input images into a single coherent generation:<ul><li>If both inputs are active, the system blends their visual and semantic context — e.g., “the man and the woman are dancing.”</li><li>If only one image is active, the model evolves that image using the new prompt — e.g., “same face and clothes, the man is riding a horse.”</li></ul>This setup enables dynamic, instruction-based image merging without losing character identity.<hr /><h3 id="lora-power-loader">🧠 LoRA Power Loader</h3>A custom LoRA Power Loader lets you stack multiple LoRAs simultaneously, each with independent strength control. You can fine-tune how each LoRA influences the output — controlling aspects like:<ul><li>Facial identity</li><li>Body type and proportions</li><li>Clothing style</li><li>Lighting or artistic detail</li></ul>This provides granular creative direction for dataset generation or character refinement.<hr /><h3 id="latent-upscale-and-restoration">🔍 Latent Upscale &amp; Restoration</h3>A built-in Latent Upscale node enhances weak or low-resolution inputs before re-generation. It restores missing structure and detail at the latent level, producing clean, high-resolution results without introducing blur or artifact noise.Use this feature to:<ul><li>Improve dataset quality</li><li>Restore bad or compressed source images</li><li>Regenerate old renders into consistent HD-quality material</li></ul><hr /><h3 id="multi-prompt-multi-step-generation">🧩 Multi-Prompt / Multi-Step Generation</h3>The workflow supports multiple prompts and generation passes, automatically rendering your character in:<ul><li>Different poses</li><li>Varied lighting and environments</li><li>Multiple angles and camera distances</li></ul>This produces a rich, consistent dataset ideal for training LoRAs or any character model requiring 360° diversity and visual coherence.<hr /><h3 id="use-cases">🎯 Use Cases</h3><ul><li>Creating training data for LoRA or DreamBooth</li><li>Generating pose/lighting variation sets for the same character</li><li>Cleaning and enhancing low-quality reference material</li><li>Building consistent visual identity across scenes</li></ul><hr /><h3 id="requirements">⚙️ Requirements</h3><ul><li>Flux-Kontext nodes (ComfyUI Flux / Florence integration)</li><li>LoRA Power Loader node</li><li>Latent Upscale node</li><li>Any Flux-compatible base model</li><li>Optional: additional LoRAs placed in <code>/models/Lora/</code></li></ul>

Flux-Kontext - Wan 2.2 Character Lora Training Image Generation

🧠 Flux-Kontext Character Creator v2.1 — Advanced LoRA TrainerThis version expands on the original character training workflow with a complete expression and motion suite for ultra-consistent LoRA training. It now captures every aspect of personality, anatomy, and lighting for professional-grade dataset generation.🌟 What’s New🎭 Full Expression RangeAdded five emotional variants to teach facial consistency and versatility:<ul><li>Neutral</li><li>Happy (natural smile)</li><li>Angry / focused</li><li>Sad / contemplative</li><li>Surprised / shocked</li></ul>Each expression maintains identical clothing, proportions, and lighting to preserve identity across moods.🏃 Dynamic PosesIntroduced movement and gesture awareness to improve body and limb accuracy:<ul><li>Walking / mid-step</li><li>Reaching or waving</li><li>Action / tension poses (arms extended or turning)</li></ul>These help LoRAs generalize character movement and physical realism.💡 Lighting ControlThree distinct lighting setups expand realism and visual depth:<ul><li>Soft studio lighting (neutral tone)</li><li>Backlit silhouette (outline training)</li><li>Harsh midday sunlight (contrast and detail recovery)</li></ul>This ensures your LoRA adapts to multiple illumination conditions without losing likeness.📝 Improved Caption SystemThe auto-caption sidecar <code>.txt</code> files have been upgraded to include full prompt context:<pre><code>&lt;lora_name&gt;, smiling, close-up portrait, same outfit, white background 
&lt;lora_name&gt;, inside dark castle, candlelight, scared expression, dramatic moonlight 
&lt;lora_name&gt;, walking, outdoor trail, daylight, relaxed expression 
...
</code></pre>Every generated image now includes a precise training tag with emotion, pose, and lighting data — ready for LoRA training pipelines.🧩 Ideal Use CaseDesigned for character LoRA creation, pose reference generation, and multi-scene identity training, this workflow automates complex dataset creation in a single run — all while maintaining perfect visual and contextual consistency.

Files with SHA256: 707658c52d0c31b70fb639c18b804a0eef40fb8826585cd761eaa8f4cf724683