Лунапопик Альтмана

Превращаемся в вервольфа в Seedance 2.0 Отправляем промпт в бот со своим фото, модель Seedance 2.0, 15 секунд: The main character matches the exact appearance, style, and features from the uploaded reference image. Maintain visual consistency throughout the video. A single continuous 15-second handheld horror shot at a seaside dinner table at night, set on a wooden beach deck beside the sand. Warm string lights sway in the ocean wind, candles flicker across the table, and dark waves roll behind the guests. Photorealistic practical-effects realism, frantic reactive camerawork, natural motion blur, no cuts. 1. 0–1.3s: The main character laughs with friends around an elegant beachside dinner table. The character suddenly freezes mid-laugh, eyes widening as a low unnatural growl emerges. 2. 1.3–2.6s: The character's head drops downward. Beast-like features begin to emerge, shoulders broaden, and fingers curl as bones shift. Friends fall silent and push their chairs back through the sand. 3. 2.6–3.9s: The character convulses against the table, overturning plates, glasses, candles, and dishes. Dark fur rapidly spreads across the arms and face while the outfit stretches with the body shift. 4. 4.3.9–5.1s: Spine arches, limbs elongate, and face pushes forward into a snarling wolf muzzle, completing a transformation into a large humanoid werewolf figure. 5. 5.1–6.4s: The transformed character drops onto all fours and charges between the chairs as guests scatter across the sand. 6. 6.4–7.6s: A fleeing guest collides with a drink cart, overturning it. Bottles, glasses, and ice scatter across the wooden deck. 7. 7.6–8.9s: The creature rapidly scrambles over the overturned dinner table toward a panicked guest near the beach railing. 8. 8.9–10.1s: The cornered guest swings a wooden chair in defense, deflecting the creature backward into the hanging string lights, sending sparks overhead. 9. 10.1–11.4s: The creature instantly recovers and leaps forward again. Another guest throws metal buckets to distract it. 10. 11.4–12.6s: The guest ducks beneath the leap and runs along the railing as moonlit waves crash behind. 11. 12.6–14s: The creature slams claws onto the railing deck. The guest slips free and dives toward the beachfront restaurant entrance. 12. 14–15s: The glass-and-wood door slams shut. The creature strikes the door frame, then turns and snarls directly at the trembling camera beneath violently swaying lights. Visual direction: Grounded seaside horror, believable physical transformation, wet realistic fur, consistent character continuity, strong ocean wind, blowing tablecloths, crashing waves, rattling dishes, scraping claws, panicked atmosphere, and distant thunder. Realistic cinematic lighting, 24fps, 16:9. Negative Prompt: Subtitles, text, watermarks, duplicate people, distorted anatomy, cartoon style, oversaturated colors, CGI look, bad physics. Бот для генераций видео: @videovvbot #seedance

Фото с рисунком себя на стене Прикрепляем своё фото к промпту в боте, модель GPT Image 2. Create a premium editorial portrait featuring the uploaded person standing beside a playful ink-twin doodle version of themselves. FORMAT LOCK Vertical 4:5 composition. Full-body view. Both figures fully visible. Side-by-side layout. Clean minimalist composition. IDENTITY LOCK Use the uploaded image as the only identity reference. Preserve the exact face, facial features, eye shape, nose, lips, skin tone, hairstyle, accessories, outfit, shoes, colors, and overall appearance. Do not invent or modify any details. REAL SUBJECT — STRICT POSE PRESERVATION The real person must keep the exact same body pose, posture, hand position, leg position, head angle, facial expression direction, and overall silhouette from the original uploaded image. Do not change the pose. Do not create a new pose. Do not mirror the pose. Do not stylize the body position. Recreate the person as a realistic photograph with the identical stance and gesture from the reference image. Keep the person completely photorealistic with natural skin texture, realistic proportions, authentic lighting, and accurate anatomy. Match the outfit exactly from the reference image. INK TWIN Create a rough black ink, charcoal, or marker doodle version of the same person drawn directly on the wall beside them. The doodle should be recognizable through the same hairstyle, swimsuit silhouette, accessories, and general pose from the real person, but it should look like a hand-drawn sketch, not a realistic duplicate. The ink twin may exaggerate the personality and add playful cartoon energy while keeping the same basic pose: loose sketch lines rough marker strokes imperfect hand-drawn style comic details such as stars, hearts, motion lines, small scribbles BACKGROUND & LIGHTING Replace the original environment with a clean white, cream, or light gray wall. Soft natural daylight-style lighting. Even illumination. Minimal shadows. Luxury editorial atmosphere. STYLE Luxury fashion editorial photography. Photorealistic person combined with hand-drawn illustration. Contemporary magazine campaign. Premium mixed-media artwork. Minimalist modern social media aesthetic. CAMERA Professional editorial photography. Full-body fashion portrait. 85mm lens. Shallow depth of field. Natural skin detail. High-end fashion campaign lighting. OUTPUT Ultra-detailed 8K quality. Sharp focus. Premium editorial photography. Clean composition. Fun, stylish, realistic mixed-media artwork. NEGATIVE PROMPT Do not change the person's pose. Do not change hand placement. Do not change leg position. Do not create a different gesture. Do not invent appearance. Do not change outfit. Do not create a realistic second person. No mirrored poses. No extra limbs. No distorted hands. No cartoon body. No text. No logos. No watermarks. No low quality. Бот для генераций фото: @img_aibot #gptimage

Создаем стильную фотку в стиле Gorillaz Прикрепляем своё фото к промпту в боте, модель Nano Banana Pro. Промпт в комментариях. Бот для генераций фото: @img_aibot #nanobanana

Делаем красивую рекламу своего продукта Как создать такое видео с любой вещью: 1) Переходим в бота, жмем /start -> выбрать нейросеть -> Nano banana Pro, разрешение 16:9 и отправляем промпт (в нем меняем название вашего продукта): Generate a hyperrealistic aerial photograph taken from inside a helicopter, looking out over midtown Manhattan on a clear bright day. Part of the helicopter's dark window frame and open door edge is visible in the left foreground, slightly out of focus, framing the shot. Beyond it, the city opens up: a wide Manhattan avenue running between dense blocks of skyscrapers, yellow cabs and traffic flowing along the street, crosswalks and tiny pedestrians on the sidewalks far below. Standing upright in the middle of the avenue intersection is a colossal version of the product from the first attached image, scaled to the height of the surrounding buildings, roughly twenty storeys tall. It is planted directly on the asphalt at the center of the intersection, dwarfing the traffic circling its base, casting a long hard shadow that stretches across the street and climbs partway up the building facades behind it. Match the first attached image precisely: same colors, same materials, same finish, same proportions, sharp and clean at monumental scale. Use the logo from the second attached image exactly as it appears, reproduced faithfully with no changes to its shape, lettering or proportions, rendered on the product's surface in its correct position, sharp and legible. Do not redraw, restyle or invent any lettering. Shot from roughly forty storeys up at a downward angle, wide-angle lens, deep focus, slight atmospheric haze in the distance, faint lens vignette. Bright midday sunlight from the upper right. Natural documentary aerial photography, shot handheld from a helicopter, subtle motion in the frame. Hyper-detailed, photorealistic. Luxury [PRODUCT TYPE] editorial. 2) Переходим в бот, /start -> выбор модели -> Seedance 2.0 (10s) 16:9 -> отправляем получившееся фото из 1 пункта. Animate the start frame as one continuous aerial shot, no cuts. The helicopter drifts slowly sideways and banks gently, so the giant [PRODUCT] in the street below shifts against the buildings behind it with real parallax. The camera stays handheld, with the faint constant vibration of a running rotor. Far below, traffic keeps moving through the intersection and around the base of the product, and the shadow it casts creeps almost imperceptibly across the avenue. The product itself stays completely still and solid, like a real installation. Keep it identical to the start frame: same colors, materials, finish and proportions. Keep the logo sharp and legible. Real documentary helicopter footage, bright midday daylight, natural city sound and rotor noise, no music. Бот для генераций фото: @img_aibot #nanobanana Бот для генераций видео: @videovvbot #seedance

Генерируем свои фото в разных эпохах Прикрепляем своё фото к промпту в боте, модель Nano Banana Pro: Create a hyper-realistic timeline collage of the same subject (use the uploaded reference photo for facial identity) portrayed across 15 decades, from the 1880s to the 2020s. Arrange all images in a clean 3×5 or 5×3 grid with uniform square/portrait framing, even spacing, and a neutral museum-archive background between cells. Label each square with its decade in a minimal serif or archival typewriter font, bottom-left corner. Consistency requirements: Preserve the exact facial structure, bone structure, eye shape, and proportions of the reference subject in every panel — only fashion, hairstyle, color grading, lighting, and photographic technique should change to match the era. Each portrait should look like an authentic photograph taken in that decade, not a modern photo with a filter. Decade-by-decade breakdown DecadeStylingPhotographic treatment1880sTraditional regional/cultural attire appropriate to the subject's heritageSepia tone, tintype/daguerreotype look, stiff studio pose1890sFormal suit or high-collar dress, bow tie or laceVintage studio lighting, soft vignette1900sColonial-era outfit, wide-brim hatWarm brown monochrome tones1910sMilitary-inspired or Edwardian formalwear, slicked/pinned hairGrainy silver-gelatin print look1920sEuropean tailoring, beret or cloche hatSoft film grain, slightly Бот для генераций фото: @img_aibot #nanobanana

Генерируем любовную открытку Love is Прикрепляем фото парня и девушки к промпту в боте, модель Nano Banana Pro: Используй прикреплённое референс-фото женщины и мужчины как эталон идентичности. Сохрани внешность и черты лица 1:1: форма лица, пропорции, глаза, нос, губы, возраст, тон кожи, текстура кожи с порами, естественная асимметрия. Не менять причёску/цвет волос (как на референсе). Без бьюти-ретуши, без “кукольной” кожи, без стилизации. Гиперреалистичное, детализированное фото винтажного вкладыша от жевательной резинки “Love is…”, слегка мультяшная иллюстрация милой пары по референсу, где девушка сидит с идеальным “я не злюсь” лицом и демонстративно отвернулась, а парень с виноватой улыбкой держит в руках шоколадку как мирный договор, забавная надпись о любви: “Любовь это… понимать, что “всё нормально” — это вообще не нормально”. Изображение должно выглядеть как реальное фото физического вкладыша на белом столе, вокруг на столе разбросаны конфеты в упаковке love is…, открытки сердечки, маленькие сердечки, ленты. Высокое разрешение, профессиональное студийное освещение, малая глубина резкости, тема Дня святого Валентина. Бот для генераций фото: @img_aibot #nanobanana

Становимся героем манги Отправляем своё фото с промптом ниже в бот, модель Gemini Omni, длительность 10 секунд: 10-second vertical cinematic video, 9:16, ultra-realistic premium live-action. A person faces the camera in a modern indoor room with warm natural ambient light. CORE VISUAL LAW Their fingertips control a dynamic portal between photorealistic reality and a black-and-white manga universe in real time. There is NO physical object between the hands — the opening exists only because of finger positions. Everything OUTSIDE the hand-defined shape stays photorealistic; everything INSIDE becomes a perfectly aligned manga version of the same face, body and room. The manga layer stays spatially locked to the real scene beneath, and the boundary follows the fingertips precisely and continuously. 0-2s — THE SLIT Medium close-up, frontal. The person looks into camera and raises both hands, thumbs and index fingers forming an extremely thin horizontal slit across the eyes. The instant the fingertips align, only the space between them turns manga — only the eyes in the slit are illustrated, while forehead, nose, mouth, hands, body and room stay photorealistic. The fingers move slightly up and down and the manga strip follows perfectly, like a scanner revealing another dimension. 2-4s — SHAPE MORPHING The hands pull apart and the slit expands into a rectangular portal, revealing more of the illustrated face. Without releasing the gesture, fingertips move independently and continuously morph it: rectangle to trapezoid to sharp triangle. NO CUT, NO physical frame, NO floating screen — the boundary stretches and bends exactly with the fingertips. As the triangle shrinks, only one manga eye remains. The person slides it across the face, transforming only the areas it passes over. 4-6s — IMPOSSIBLE GEOMETRY The person rotates both wrists in a fast elegant motion and the triangle bends into a circular opening between the hands, fingers curving around its invisible boundary. Inside it, face and room appear as detailed manga artwork. The window moves across eye, nose, mouth, then the whole face — reality becomes manga inside the portal and returns to photorealism right after it passes. The person pulls the circle sideways, stretching it into an irregular organic shape like an elastic dimensional membrane, the manga universe stretching in correct perspective. 6-8s — REALITY TEAR The person grabs opposite sides of the portal and pulls outward. The manga opening rapidly grows beyond the face toward all four screen edges, briefly consuming the entire frame. The WHOLE PERSON and ROOM become a detailed black-and-white manga universe: bold ink contours, cross-hatching, screentone shadows, speed lines, dramatic hand-drawn perspective. Identity, exact position and movement remain unchanged. 8-9s — BREAK THE COMIC WORLD Inside the full manga universe, the person looks into camera and pinches the air with thumb and index finger. A tiny photorealistic hole appears between the fingertips and they pull it apart. Real color and photographic texture break through as the manga world folds back toward the edges like an elastic membrane. 9-10s — SNAP BACK TO REALITY The person snaps their fingers and the remaining manga universe instantly collapses toward the snapping fingertips: full screen to rectangle to triangle to tiny circle to glowing point, then gone. Everything is fully photorealistic again. The person holds still for the final beat with a subtle confident smile. Бот для генераций видео: @videovvbot #geminiomni

Делаем фото в стиле бумажной открытки Прикрепляем своё фото к промпту в боте, модель Nano Banana Pro: Transform my uploaded reference photo into a beautiful handcrafted 3D layered paper cut illustration while perfectly preserving the person's identity, facial features, hairstyle, expression and pose. Create the entire artwork using stacked paper layers with realistic cardstock textures, precise paper cut edges, visible paper thickness and soft dimensional shadows between each layer. The character should have soft expressive eyes, simplified yet recognizable facial features, layered paper hair strands and smooth handcrafted details. Build the entire environment from paper including mountains, trees, flowers, clouds, buildings and decorative plants. Use soft earthy colors including cream, beige, sage green, olive green, pastel blue, peach and muted brown. Everything should look like an expensive handcrafted paper artwork photographed from the front with soft natural daylight, subtle shadows, premium craftsmanship, highly detailed layered composition and elegant framing. Бот для генераций фото: @img_aibot #nanobanana

Летаем на драконе Отправляем промпт в бот, модель Seedance 2.0, 15 секунд: Create a 15-second vertical 9:16, 24fps, 8K-quality photorealistic live-action fantasy sequence from a first-person rider POV. A massive original obsidian-crimson dragon with individually detailed scales, subtle dark mottling, molten-gold eyes, volcanic-black swept-back horns, blade-like dorsal fins, burgundy-red translucent wings, and faint ember-like fissures across its body. The rider wears black leather gloves, dark-green sleeves, and weathered leather bracers, holding heavy chain reins from a worn brown studded saddle. Start on a wet emerald coastal cliff. The dragon is already poised to move. The rider gives the dragon two quick pats on the neck, and the dragon immediately responds — it surges forward, takes just a few powerful running steps across the grass, and leaps off the cliff. It drops fast toward the grey-green ocean, then powerfully opens its huge wings with realistic membrane movement and catches the air. Continue with a fast, low flight just above the ocean surface, close enough for sea spray to hit the lens. End with a dramatic high-speed bank around a tall sea stack, emphasizing speed, weight, and momentum. Use physically accurate anatomy, gravity, wing aerodynamics, water physics, scale deformation, wind, and environmental interaction. Camera must feel naturally mounted on the rider: realistic first-person motion, subtle handheld vibration, motion blur, focus breathing, and lens droplets. Soft overcast lighting, HDR, restrained film grain, natural cinematic color grading. Make it indistinguishable from expensive real live-action cinema. Audio: no music, no dialogue, no voice. Only realistic SFX/ASMR — dragon breathing, scale movement, leather creaks, chain reins, the rider’s pats on the neck, claws scraping wet soil, running impact, wind rush, wingbeats, ocean spray, and one deep natural dragon roar during flight or the sharp bank. Avoid: AI look, cartoon, anime, game CGI, plastic textures, fake physics, neon glow, oversaturation, distorted anatomy, extra limbs, text, or watermark. Бот для генераций видео: @videovvbot #seedance

Строим дом в Gemini Omni Отправляем фото дома с промптом ниже в бот, модель Gemini Omni, длительность 10 секунд: Create a 10-second cinematic vertical (9:16) miniature repair timelapse based on the uploaded reference image. The image represents the object — vehicle, machine, gadget, furniture, building or any item — that needs repair. Tiny workers inspect, repair, restore, clean and replace damaged parts, bringing it back to brand-new factory condition. It should feel like a premium miniature repair documentary filmed inside a handcrafted scale-model world. REFERENCE LOCK (highest priority) Analyze the uploaded image first and identify: what the object is, its purpose, materials, engineering, structure, damaged or worn areas, and the most realistic repair method. The image is the single source of truth. Maintain its shape, size, design, proportions, body style, wheels, windows, doors, logos, structural details and design language. Do NOT redesign or change the object's identity. RESTORATION RULE The finished object must look brand new, as if it just left the factory or showroom. Repair or remove scratches, dents, rust, dirt, dust, broken parts, cracks, paint damage and missing components. Replace damaged parts with brand-new matching parts and restore premium surface quality. VEHICLES (highest priority) If the object is a vehicle, after repairs it receives: a full professional repaint in the same original color as the uploaded vehicle, deep glossy showroom paint, crystal-clear reflections, fresh clear coat, brand-new polished alloy wheels, clean tires, spotless windows, chrome polishing, bright headlights, clean taillights, detailed interior cleaning, engine detailing, final ceramic coating and professional polishing. It must look like a brand-new showroom model while keeping the same make, model, body shape and design as the uploaded image. REPAIR PLANNING Automatically choose the most realistic repair process for the uploaded object. Vehicles — miniature mechanics, hydraulic lifts, wrenches, power drills, welding tools, replacement panels, engine tools, tire-changing equipment, spray booth, paint guns, buffing and polishing machines, ceramic coating gear. Workers repair and replace damaged parts, weld, repaint panels, polish, detail and run a final quality inspection. Electronics — precision screwdrivers, soldering stations, tiny robotic arms, circuit repair tools, magnifiers, replacement chips, cleaning brushes. Workers repair internal components, clean the device, replace damaged parts, reassemble and test it. Buildings — miniature construction workers, scaffolding, cement mixers, bricklaying tools, glass installation, paint rollers, repair cranes. Workers fix structural damage, replace broken materials, repaint, clean and restore it. Furniture — carpenters, hammers, drills, glue, sandpaper, brushes, wood polish. Workers repair joints, replace damaged wood, repaint and polish to a premium finish. Machinery — engineers, hydraulic tools, cranes, replacement gears, lubrication systems, industrial repair equipment. Any other object — determine the most realistic process from its material, engineering, manufacturing and craftsmanship, using only appropriate real-world tools and techniques. Бот для генераций видео: @videovvbot #geminiomni