Floating Kebab Boat: Candid Photo Then Animate Into a Cinematic Video

First, you need to generate a realistic frame, and GPT Image 2.5 (Flare) is great for that along with this prompt: Authentic candid smartphone photo, sunny tourist bay near a beach. Soft natural daylight, unpolished real phone snapshot, slight phone compression, no text, no watermarks, no UI overlays. CAMERA (critical - low and flat): phone held just above the water surface - almost water-level, only slightly above the ripples - shooting straight ahead horizontally. Flat frontal angle, camera parallel to the sea. NO high angle, NO looking down, NO picada, NO elevated viewpoint, NO drone. Very low POV as if the photographer is crouching or wading with the phone barely clear of the water, aiming straight at the boat. Because the camera is so low: the turquoise water fills the lower half of the frame; people's legs and waists loom large mid-frame; their heads sit high in the frame; the boat stern is seen frontally at that same low height. Frontal view of a small motorboat / open lancha at anchor in shallow turquoise water, stern facing camera. On the stern: vertical rotating doner kebab spit with cooked meat, compact serving setup - cutting board, knives, flatbreads/pita stack, sauce bottles, paper wraps, napkins - floating kebab stand. A vendor working the spit. Tourists in swimsuits standing waist-deep in a loose queue between camera and boat, empty-handed, waiting to be served - still waiting to eat, no wraps, no food in their hands. Swimsuits, sunglasses, vacation vibe. Nearby other small pleasure boats. Background: sandy beach with sunbathers and umbrellas, low coastal buildings, clear sky. Mediterranean holiday coast. Photoreal phone photo, everyday depth of field. Once you lock it, it's time to animate it with Seedance 2.5 and a prompt like this one: [STYLE + CAMERA + ATMOSPHERE] Photoreal live-action tourist phone video. Continuous single take, handheld from low in the water - almost water-level, slight above the surface, flat frontal, camera parallel to the sea. Digital zoom in / zoom out only (no optical cuts, no hard edits). Soft natural midday sun, clear turquoise bay, real water ripples, real skin, real boats. People act natural; a few glance at the lens for a beat then look away. Diegetic sound only: water splash, knife on spit, voices, whistle, soft cat landing. [IMAGE REFERENCES] [ref_image] = locked master still. Keep exact boat, kebab spit, canopy, vendor, bay, buildings, water color, and the queue of tourists in swimsuits as the opening frame. Same low frontal POV. Add only what the timeline requires (cash exchange, large jellyfish left of frame, black cat on the boat). [CAST + PROPS - LOCKED] Vendor on the stern: white tee, dark apron, white/dark cap, cutting doner from the vertical spit. Customers in the water empty-handed at start, waiting in a loose queue. One customer pays with a paper bill. Next in line: blonde woman in swimsuit. Black cat appears from the rear of the boat later. Large real jellyfish in the water on camera-left. [TIMELINE - 30s, ACTION ALREADY ON] 0-1.5s: Already mid-service. Digital zoom pushes in on the vendor's hands - knife shaving meat from the spinning spit onto a warm flatbread on the board. Water lapping at lens height. A tourist in the queue glances at camera then back to the boat. 1.5-3s: Zoom pulls back a touch. Vendor finishes the wrap, leans over the stern, hands it to the waiting customer in the water. Hands empty until they receive the wrap - exchange already happening. 3-5s: Customer holds the wrap, pays with a paper banknote. Vendor takes the bill, stuffs it into his apron pocket in one practiced motion. Mild digital zoom on the pocket. 5-7s: Vendor looks to the next person - the blonde woman in the queue - raises his eyebrows in a clear "you're next" gesture. She answers with a short spoken line (mouth moves, casual). Another tourist glances at camera. 7-10s: He goes into overdrive - knife flying on the spit, meat shreds falling fast onto a new flatbread, hands blurring. Digital zoom in-out on the spit and his face. Queue watches, still waiting, empty-handed. 10-12s: Still normal tourist chaos - ripples, boat rock, distant beach umbrellas, another dinghy drifting. Continuous handheld sway. Everything still feels like a real holiday clip. 12-14s: Sudden shift. People in the water flinch, turn, start scrambling away from camera-left. Handheld whip-pan LEFT into the open water. 14-16s: A large real jellyfish drifts and pulses near the surface on the left - clear gelatin body, tentacles trailing in turquoise water. Tourists push back, splash, clear the left side. 16-18s: Camera swings back RIGHT to the kebab boat. Vendor cups his mouth and WHISTLES sharp and short. From the rear of the boat a black cat vaults into view, lands on the stern edge facing the water. 18-20s: Cat freezes one second on the gunwale - low crouch, eyes locked on the jellyfish - then launches into the water in a clean dive. 20-23s: Cat bites into the jellyfish, thrashing once, then springs back UP onto the boat deck in one leap, jelly still in its mouth. Splash, wet fur, boat rocks. 23-26s: Vendor grabs the jellyfish from the cat and tosses it onto the boat deck floor - it lands wet on the planks and stays there. He pets the wet black cat hard and affectionate. 26-30s: He tears a piece of kebab meat from the spit and feeds it to the cat. Cat takes it. Soft digital zoom out to a wider hold: boat, spit, cat eating, jellyfish on the deck floor, leftover queue edge, sunny bay. Handheld settles. Continuous take ends. [STYLE & QUALITY BOOSTERS] Real phone holiday video, continuous handheld low-water frontal POV, digital zoom only, photoreal water and skin, doner service then jellyfish scare then black-cat hero recover, vendor tosses jellyfish on deck floor then pets and feeds the cat, coherent physics, natural tourist behavior with occasional lens glances.