How to Create an Empire State Building Spire Kiss Video With AI
Turn two portrait photos into a dramatic news-helicopter style couple video inspired by the viral Empire State Building spire story. The final clip feels like distant telephoto footage from a circling helicopter: two black-clad figures on the antenna mast, a huge black banner whipping in the wind, masks coming off, and a kiss at the top of New York City.
What you will make
This effect is designed for two-person reference-to-video generation. It uses the uploaded photos for identity, then places the couple into a cinematic high-altitude scene with black sleeveless outfits, face coverings, a wind-blown banner, helicopter shake, and a continuous telephoto zoom.
Pixnova source example
This tutorial uses a Pixnova-generated character preview as the visual example for the filter cover and result video.
Pixnova task ID:
c338a878-0f96-4747-868f-83138fa54b64
Use the video above as the creative reference: the mast-top placement, the black sleeveless outfits, the banner scale, and the final kiss are the key details to reproduce.
Best photos to upload
Upload two clear adult identity photos. The first photo should be the man, and the second photo should be the woman. Clear faces, natural lighting, and visible hair shape help the AI preserve identity after the masks are removed.
News inspiration
On July 1, 2026, Angela Nikolau and Ivan Beerkus / Ivan Kuznetsov climbed to the top of the Empire State Building antenna structure in New York City. Reports from ABC News, The Guardian, and The National Desk described the pair reaching the spire, wearing dark clothing, unfurling a black peace-message banner, and later being arrested.
This Vidream effect is a fictional AI recreation inspired by the visual language of that news moment. It is not a real stunt tutorial and should not encourage climbing, trespassing, or unsafe behavior.
Step 1: Choose the Empire Spire Kiss effect
Open the Empire Spire Kiss video effect. It is built for two uploaded photos and uses a 15-second vertical 9:16 HappyHorse reference-to-video workflow.
The effect focuses on five visual anchors:
- The absolute top of the Empire State Building antenna mast
- Two black sleeveless climbing outfits
- Black face coverings removed near the end
- A huge black banner attached to the woman and pulled open by wind
- A distant helicopter telephoto camera that keeps orbiting
Step 2: Upload two photos
Use one clear photo for each person. Avoid sunglasses, heavy face filters, low-resolution screenshots, or photos where the face is tiny. Because the scene includes masks at first, the most important identity moment happens after the masks come off.
For the best result:
- Use adult portraits with visible faces
- Keep lighting natural and not too dark
- Use photos with stable hairstyles
- Upload the man first and the woman second
Step 3: Generate the video
The effect creates a continuous helicopter-style zoom. The shot starts far away, moves in with telephoto compression, and stays distant enough to keep the antenna mast, banner, and both people visible. The couple should not look into the camera. The emotional payoff is private: they look at each other, remove their masks, and kiss while the banner keeps flapping in the wind.
If the first result misses the banner or kiss, regenerate once or twice. This scene asks the model to handle multiple hard elements: a tiny mast-top location, two identities, face coverings, fabric physics, and a kiss. Results are strongest when the banner and kiss are emphasized in the prompt.
Why this prompt works
The prompt avoids asking for too many story beats. There is no proposal, no ring, and no kneeling. That keeps the model focused on the three most important actions: banner, mask reveal, and kiss.
It also avoids close-up beauty shots. A close-up can make the clip feel like a staged music video. The better look is imperfect broadcast footage: long-lens shake, compression noise, haze, and a circling helicopter perspective.
Full prompt
Create a realistic 15-second vertical 9:16 news-helicopter video using the uploaded couple reference photos. After the masks are removed, preserve both identities clearly.
MUST HAVE: a huge black banner attached to the woman's body/arm/waist, unfurling in strong wind; both people remove black face coverings; they never look at the camera; they look only at each other and kiss. No proposal, no ring, no kneeling.
Scene: the couple stands at the absolute top of the Empire State Building antenna mast, on the tiny metal mast cap above the main spire. They are beside the vertical antenna mast at the highest visible point. Not a rooftop, not an observation deck, not a wide platform.
Wardrobe: both wear black sleeveless climbing outfits. Man on the left: fitted black sleeveless top, black technical pants, black shoes, black face covering, visible bare arms. Woman on the right: black sleeveless crop top or sleeveless technical top, black climbing pants, black shoes, black face covering, visible shoulders and arms. Matte black, tight, functional, windblown.
Banner: the woman carries a huge black fabric banner attached to or draped from her body, arm, or waist. It is NOT tied to a railing. Strong wind pulls it outward and downward from her side. The banner is long, curved, twisted, rippling hard, with large white block letters such as "POWER OF LOVE / WORLD PEACE". The banner must stay visible throughout the video.
Camera: one continuous shot from a helicopter circling the antenna mast. Long telephoto news lens, 600mm-800mm. No hard cut. Start far away, then continuous optical zoom-in while the helicopter keeps orbiting. Keep long-lens shake, aircraft vibration, focus breathing, haze, compression noise, slightly soft broadcast quality. Even after zooming in, stay medium-long telephoto, not close-up. Keep the mast, tiny top structure, banner, and both people visible.
0-5s: distant shaky helicopter shot. Two tiny black figures stand on the very top mast cap. The woman's black banner is already visible, blowing outward from her body.
5-10s: continuous telephoto zoom-in, no cut. The helicopter keeps orbiting. The mast and couple grow larger; the banner whips hard in the wind.
10-15s: medium-long telephoto view, still shaky. Both remove their black face coverings, reveal the reference faces, look only at each other, lean close, and kiss. The banner keeps flapping beside them.
Audio: no music, no romantic soundtrack, no score. Only helicopter rotor rumble, aircraft vibration, strong wind, banner flapping, distant city ambience, clothing rustle.
Avoid: missing banner, missing kiss, proposal, ring, kneeling, looking at camera, selfie pose, wide rooftop, observation deck, broad platform, railing-mounted banner, long sleeves, jackets, clean 4K MV look, hard cut, face close-up, drone-smooth footage, police, falling, panic, climbing tutorial, distorted faces, identity mixing, extra fingers, watermark.
Pro tip
For the most reliable output, start from a preview image where the woman already carries a large black banner and both people are already near the antenna mast. HappyHorse is strongest when it animates a clear starting composition instead of inventing every prop from text alone.