A dynamic cinematic dark fantasy scene, one continuous 10-second shot with a perfectly seamless loop. The camera remains almost stationary, with only a very subtle smooth side-to-side cinematic drift. No zoom, no push-in, no pull-back, and no sudden camera movement. The demon witch flirts directly with the viewer while preserving her original pose and appearance. She slowly tilts her head, slightly lowers one shoulder, gently arches her back, and gives the camera a playful, confident half-smile. Her glowing eyes remain locked on the viewer. She briefly looks from under her eyelashes, slowly blinks, subtly raises her chin, and then smoothly returns to her original expression and pose. Her movements are seductive, elegant, controlled, and natural rather than exaggerated. Both hands remain naturally resting on the stone altar. Her long black hair sways around her face and shoulders in the hot wind. The loose fabric of her dark dress ripples gently around her body and legs. The flames behind her rise and curl rhythmically, creating moving orange reflections across her skin, horns, hair, dress, and the carved altar. Glowing embers spiral through the air, smoke rolls slowly between the tombstones, dry autumn leaves circle through the scene, candle flames flicker, and the raven turns its head before returning to its original position. All character and environmental movements form one smooth synchronized cycle lasting exactly 10 seconds. During the final seconds, the witch gently returns her head, shoulders, expression, hair, dress, and gaze to their precise starting positions. At 10.0 seconds, the entire composition must perfectly match the first frame for an invisible seamless loop. Preserve the woman’s face, identity, anatomy, body proportions, hands, horns, clothing, and original illustration style. Keep the “Cinder Witch” title completely motionless, sharp, undistorted, and fully readable. No cuts, no transitions, no new objects, no added characters.
Parameters used to generate this content
The node graph used to generate this content — pan and zoom to explore, or download it to run in ComfyUI.
AI models used to generate this content