Snow Globe Miniature McDonald's Rose Product Video Prompt

Here's the thing.

@Soaima_Ai opens with two model names stacked together, Seedance 2.0 and GPT image 2, then the label Video Prompt, then the full prompt. The pairing tells you the pipeline. Generate the still first on GPT image 2, then hand that still to Seedance 2.0 as the reference frame and let this prompt animate it.

The reference frame is a product-shot-style still. A clear glass snow globe with a small gold loop at the top, holding a miniature McDonald's restaurant inside. Red roof, golden arches, the McDonald's name in white lettering across the fascia, tiny interior lights glowing warm, a micro-scale parking apron in front. The globe sits centered on a dusty-pink surface and is ringed at its base by a dense garland of fresh roses in pink, white and cream, with baby's breath, lavender sprigs and eucalyptus leaves filling out the arrangement. Scattered petals lie on the pink surface around it, and the background is the same pink, soft and out of focus, with more blurred roses and tiny white flowers rising out of the haze on both sides.

The prompt asks the video model for exactly one camera move. A slow push in toward the glass globe with a smooth macro dolly movement. That is the whole shot, no cuts, no orbits, no whip pans. What brings it to life is the layer of small motions stacked on top of the move. Tiny flower petals fall gently through the air and settle around the base. The flowers sway subtly, as if a soft breeze is passing through the room. Warm studio light rolls across the curved glass, throwing moving reflections and highlights as the camera gets closer.

The rule that keeps the shot premium instead of busy is the depth split. The miniature restaurant inside the globe stays sharp and detailed at all times, while the floral background stays softly blurred with elegant bokeh. One thing in focus, everything else a dream.

The prompt, as posted.

Create a dreamy, ultra-realistic cinematic product-style video using the reference image. A delicate glass snow globe containing a miniature McDonald's restaurant sits beautifully in the center, surrounded by fresh pink, white, and cream roses and tiny flowers. The camera slowly pushes in toward the glass globe with a smooth macro dolly movement. Soft warm studio lighting creates beautiful reflections and highlights on the glass. Tiny flower petals gently fall through the air and settle around the base. Add subtle movement to the flowers as if a soft breeze is passing through. The miniature restaurant remains sharp and detailed inside the glass globe while the floral background stays softly blurred with elegant bokeh. Premium dreamy atmosphere, shallow depth of field, realistic glass reflections, soft pink color palette, cinematic lighting, smooth motion, 4K, photorealistic, 24fps, vertical 9:16, no distortion, no camera shake.

A small note on the original. The author is Soaima, posting as Soaima_Ai, the same account behind the watercolor autumn prompts in this library. The post is in English, the prompt ships in full in the tweet body, and the two attached media show the pipeline. The photo is the GPT image 2 reference still, and the video is the Seedance 2.0 output. The clip plays exactly as the prompt directs, the camera pushes in slowly until the curved glass fills the frame and the interior detail dissolves into reflections, petals drifting through the pink. The template generalizes well beyond fast food, swap the miniature restaurant for any tiny scene inside the globe and the prompt holds.

At the time of writing, the post sits at 28 likes, 4 retweets, 12 bookmarks and 765 views, with 20 replies under it. Original thread here, x.com/Soaima_Ai/status/2096407257751883832