Sharing a Character Consistency Image Workflow

BigBilly823 · reddit · 2026-07-18

The author shares their first experiment with a "character consistency" workflow: generating a character via Perchance, then using Gemini and ChatGPT to refine prompts and define visual traits.

After generating multiple scenes, they imported the assets into Krita for unified fixes: using AI inpainting for details and manually adjusting swimsuit colors, minor blemishes, hands, and eyes to preserve character and outfit consistency across images. The author emphasizes this isn't a "perfect method," but rather a complete workflow suitable for beginners, inviting others to share their own approaches.

Original post →

More from Multimodal

Multimodal channel →