Image-to-image one-click nine views

Image-to-Image One-Click Nine Views is a ComfyUI workflow designed to create highly consistent 9-image sets from a single reference photo. It preserves the same person (face/body), the same outfit, and the same background + lighting, and produces natural variations only in facial expression, body pose/action, shot size, and camera angle, with a photorealistic, professional photography look.

What it does

• Analyzes the input image and automatically generates 9 structured prompts (one per view) using Qwen-VL.

• Runs image editing / i2i generation with Qwen Image Edit, using the reference for maximum consistency.

• Builds a 3x3 grid / reel from the 9 outputs.

• Saves all images and can export a ZIP for easy download/share.

Best for

• Character consistency (datasets, LoRA prep, model sheets, UGC/profile packs).

• Quickly exploring poses + expressions without changing styling or scene.

• Fast batch content for social media or visual reference.

Inputs / Outputs

• Input: 1 reference image (your base character).

• Output: 9 edited/generated images (batch) + grid/reel + optional ZIP.

Notes

• For best results: use a clean, well-lit reference with the outfit clearly visible.

• If you want less drift, strengthen the negative prompt and restrict prompts from changing lighting/background/props.

Requirements: ComfyUI + common custom nodes (prompt utilities, string-to-list, batch/grid tools) and the corresponding Qwen models (VL + Image Edit + VAE/LoRA if used).