Illustrious Character Consistency — OpenPose + Depth Workflow

A ComfyUI workflow designed for controlled character generation using an Illustrious-based SDXL checkpoint, LoRA, OpenPose and Depth ControlNet.

What this workflow does

This workflow combines a reference image with OpenPose and Depth preprocessing to help maintain the composition, pose and spatial structure of the reference while generating the final image.

The workflow also includes a LoRA stage for character/style conditioning and a two-stage generation process followed by image upscaling and tiled VAE processing.

Main Components

• Illustrious / SDXL checkpoint

• LoRA Loader

• OpenPose ControlNet

• Depth ControlNet

• Depth Anything V2 preprocessing

• OpenPose preprocessing

• Two KSampler stages

• 4x UltraSharp upscaling

• Tiled VAE encoding/decoding

• Final image preview

Workflow Structure

Reference Image → OpenPose + Depth → ControlNet Conditioning → Base Generation → Upscale → Refinement → Final Output

The workflow is organized into clearly separated sections to make the process easier to understand and modify.

Requirements

You will need the required ComfyUI custom nodes and model files used by this workflow.

The workflow currently references:

• waiIllustriousSDXL_v170.safetensors

• Reze.safetensors

• IllustriousXL_openpose.safetensors

• illustriousXLv1.1_depth_midas_fp16.safetensors

• depth_anything_v2_vitl.pth

• 4x-UltraSharp.safetensors

Replace the checkpoint and LoRA with your own compatible models if needed.

Notes

The workflow is intended for users who want greater control over pose, depth, character consistency and image composition when working with Illustrious in ComfyUI.