Реклама для бренда батончиков на 1 МЛН просмотров. Вторая часть промпта для анимации:
Используйте ChatGPT Image 2.0 и Seedance 2.0 на SpeShu.AI с промокодом
9.2–11.8 seconds — MACRO SNAP:
A large foreground chocolate fragment rapidly crosses the lens and creates an impact cut into @Image7. In extreme slow-motion macro, the exact chocolate bar snaps once into two pieces. The pieces move apart while one thick glossy caramel strand stretches naturally between them. Wafer layers, cereal spheres and chocolate flakes remain sharp and recognizable. The camera makes a short lateral tracking movement with a precise rack focus from the upper half to the lower half. One powerful crisp crunch, no repeated breaking.
11.8–15.0 seconds — FINAL PACKSHOT:
The caramel strand sweeps across the lens as a warm amber wipe. The camera smoothly pulls back and resolves into the exact final composition of @Image8.
The intact KRAKK package stands behind the bar while the broken piece settles gently in the foreground. A few crumbs fall onto the surface and the remaining particles slowly lose momentum.
Preserve the exact package design, product proportions, cross-section and lighting from @Image8. Hold the final packshot almost still for the last 1.2 seconds, with only subtle drifting particles, a soft copper glow and a very slow final camera push-in.
VISUAL STYLE:
Ultra-photorealistic premium food advertising, dark charcoal-black and molten copper palette, realistic caramel viscosity, believable chocolate physics, crisp wafer texture, shallow depth of field, controlled motion blur, dramatic side and rim lighting, polished high-end CGI combined with macro commercial photography.
SOUND DESIGN:
Dark premium electronic rhythm with deep restrained bass pulses synchronized to each transformation. Detailed chocolate cracking, sticky caramel stretching, flying crumbs, liquid chocolate movement, one powerful wafer crunch and a subtle wrapper snap. Finish with a deep clean product-impact sound. No voiceover and no dialogue.
STRICTLY AVOID:
No slideshow, no static Ken Burns effect, no slow dissolves, no random objects, no duplicate bars, no duplicate packages, no product deformation, no melting geometry, no incorrect internal layers, no altered logo, no misspelled text, no extra text, no subtitles, no watermark, no hands and no people.
Используйте ChatGPT Image 2.0 и Seedance 2.0 на SpeShu.AI с промокодом
TG15 — получите бонусом 15% к вашему пополнению.👍3💯3🤩2
Половина лета прошла, а у вас до сих пор нет фото с курорта?
Исправьте это в Фотостудии SpeShu.AI — готовые летние образы ждут вас. Просто отправьте своё фото, и сервис выдаст серию из трёх снимков в одной стилистике 🔥
Дополнительно вы можете примерить новую одежду, не ходя в магазин, и поэкспериментировать со своим образом. Лето — время обновления. Не упустите эту возможность — удивить родных и друзей короткой причёской или пирсингом 😁
Исправьте это в Фотостудии SpeShu.AI — готовые летние образы ждут вас. Просто отправьте своё фото, и сервис выдаст серию из трёх снимков в одной стилистике 🔥
Дополнительно вы можете примерить новую одежду, не ходя в магазин, и поэкспериментировать со своим образом. Лето — время обновления. Не упустите эту возможность — удивить родных и друзей короткой причёской или пирсингом 😁
🤩4👍2🤯2
⚡️ Какие обучающие посты написать для вас? Задайте вопросы в комментариях
👍3🤔2😱2
This media is not supported in your browser
VIEW IN TELEGRAM
Kimi K3 собрал игру про капибар за три промпта
Новый китайский флагман догоняет Claude Fable 5 и ChatGPT-5.6 Sol по производительности.
В бенчмарках Kimi K3 легко справляется с многоэтапными задачами, управляет агентами и пишет неубиваемый код. Контекстное окно — аж 3 МЛН токенов. Нейросеть часами «помнит» всё, о чём вы её попросили 🤯
При этом Kimi K3 в несколько раз дешевле американских лидеров рынка. Так что если хотите сэкономить на крупной задаче, это умный способ.
Сэкономьте ещё больше со SpeShu.AI — Kimi K3 уже на платформе. Получите бонус по промокоду
Новый китайский флагман догоняет Claude Fable 5 и ChatGPT-5.6 Sol по производительности.
В бенчмарках Kimi K3 легко справляется с многоэтапными задачами, управляет агентами и пишет неубиваемый код. Контекстное окно — аж 3 МЛН токенов. Нейросеть часами «помнит» всё, о чём вы её попросили 🤯
При этом Kimi K3 в несколько раз дешевле американских лидеров рынка. Так что если хотите сэкономить на крупной задаче, это умный способ.
Сэкономьте ещё больше со SpeShu.AI — Kimi K3 уже на платформе. Получите бонус по промокоду
TG15 и забудьте про дорогие подписки за $ и €.⚡5👨💻4✍3
Спорим, вы не угадаете, это реальное фото или нейросеть
Мы раздобыли промпты, которые выкручивают реализм ваших генерация на 10 000%. Добавляйте эти абзацы к основному запросу и получайте сверхдетальные изображения.
Чтобы сохранить внешность точь-в-точь:
Физически правильный свет:
Реализм лица и кожи:
Негативный, который убирает пластик:
Протестируйте с бонусом 15% по промокоду
Хотите бесплатные генерации? Подключите подписку и получите до 500 фото в подарок, а также SpeShu Claude, Gemini и ChatGPT.
Мы раздобыли промпты, которые выкручивают реализм ваших генерация на 10 000%. Добавляйте эти абзацы к основному запросу и получайте сверхдетальные изображения.
Чтобы сохранить внешность точь-в-точь:
Use Image 1 as the strict identity reference. Preserve the exact facial structure, proportions, skin tone, eyebrows, eye shape, nose, lips, jawline, hairline and recognizable natural asymmetry of the person. Do not replace, redesign, idealize or significantly beautify the face.
Физически правильный свет:
Use one physically coherent primary light source with a clear direction, supported by subtle natural bounce light from the environment. Preserve believable shadow direction, natural highlight roll-off, realistic skin translucency and accurate reflections in the eyes. Avoid flat frontal illumination and conflicting light directions.
Реализм лица и кожи:
Render physically accurate, non-uniform human skin without digital smoothing. Show irregular pores of different sizes, subtle pigmentation variation, tiny natural blemishes, faint redness, fine lines, delicate under-eye texture, natural eyelid folds, individual vellus facial hairs, baby hairs along the hairline and slight facial asymmetry. Preserve realistic skin translucency, natural oil distribution and subtle microcontrast. The lips show individual vertical lines, slightly uneven moisture and believable soft-tissue texture.
Render detailed human eyes with individual iris fibers, natural pupils, transparent corneal depth, moist tear lines, subtle scleral capillaries, individual eyelashes and physically accurate catchlights. Hair consists of separate strands of different thicknesses with natural flyaway hairs rather than a perfectly smooth artificial mass.
Негативный, который убирает пластик:
The final image must look like an unretouched RAW photograph captured with a real physical camera, not CGI or polished AI artwork. Preserve restrained color grading, believable white balance, natural contrast, subtle sensor grain and realistic optical softness. No beauty filter, skin smoothing, waxy or porcelain skin, artificial symmetry, excessive sharpening, fake HDR, oversaturated colors, repeated skin texture, painted eyes, plastic hair or CGI appearance.
Протестируйте с бонусом 15% по промокоду
TG15 на SpeShu.AI в ChatGPT Image 2.0.Хотите бесплатные генерации? Подключите подписку и получите до 500 фото в подарок, а также SpeShu Claude, Gemini и ChatGPT.
👍5⚡3🤯2
Media is too big
VIEW IN TELEGRAM
Новый залипательный тренд 😍
Как только вы поставите 10 лайков под этим постом — поделимся промптом, чтобы вы могли повторить и собрать свой первый МЛН просмотров
Как только вы поставите 10 лайков под этим постом — поделимся промптом, чтобы вы могли повторить и собрать свой первый МЛН просмотров
👍15✍3👨💻3
СПЕШУ АИ: Проводники в ИИ | ChatGPT, Claude, Gemini и ещё 300+ нейронок
Новый залипательный тренд 😍 Как только вы поставите 10 лайков под этим постом — поделимся промптом, чтобы вы могли повторить и собрать свой первый МЛН просмотров
Спасибо за лайки — публикуем 1-ю часть промпта
REALISM AND CAMERA:
Immersive human-eye-level action-camera perspective, broad natural depth of field, physically coherent forward momentum, realistic acceleration and deceleration, consistent warm sunlight from the upper left, believable reflections, shadows, scale and environmental motion. Motion blur affects only elements moving relative to the camera, never the entire image. No artificial stabilization warping or exaggerated fisheye distortion.
SOUND:
Generate synchronized environmental sound: tropical birds, moving leaves and waterfall ambience at the beginning; a rapidly rising wind rush during acceleration; water spray over the river; subtle city ambience between the buildings; then a smooth descending whoosh and a quiet mechanical tape-lock click when stopping at the entrance. No dialogue, narration or vocals.
STRICTLY AVOID:
Any camera cuts, new viewing angles, separate shots, frozen slideshow frames, visible morphing, warped architecture, flickering glass, changing hands, extra fingers, extra arms, duplicated tape measures, loose or broken tape, disappearing yellow line, camera rotation, collision with the entrance, added captions, additional logos, misspelled branding or distorted text. Finish on @Image2 exactly.
2-я часть промпта для тренда про рулетку
Use @Image1 as the exact opening frame and @Image2 as the exact final frame.
Create one continuous 15-second ultra-photorealistic vertical 9:16 first-person POV journey. This must be a single uninterrupted forward camera movement with absolutely no cuts, montage, crossfades, dissolves, teleportation or slideshow behavior. The viewer physically travels through one continuous environment while following the yellow measuring tape.
STRICT POV AND OBJECT LOCK:
Preserve the exact same adult male hands, forearms, skin tone, arm hair, anatomy, finger placement, natural grip, black tape-measure housing and yellow measuring tape from @Image1 throughout the entire video. The camera, arms and tape-measure housing travel forward together as one unified first-person rig. Keep the housing anchored in the lower center of the frame. The hands remain stable and recognizable without changing shape, age or position.
The yellow measuring tape must remain one real, narrow, flat metal tape with black measurement markings. It stays perfectly straight, taut and centered, continuously guiding the camera forward. It must never split, disappear, bend sideways, become a glowing beam, turn into a road or change into another object.
CONTINUOUS MOTION:
0–2 seconds:
Begin by holding @Image1 almost exactly still. The jungle feels alive: tropical leaves move gently in the humid breeze, sunlight flickers naturally through the canopy, and water flows down the waterfall on the right. The hands make only a tiny natural adjustment around the housing. The yellow tape lightly vibrates and then becomes completely taut, creating anticipation before acceleration.
2–5.5 seconds:
The camera, hands and tape measure suddenly accelerate forward together along the yellow tape. Travel rapidly down the narrow jungle path, passing wet leaves, roots, hanging vines and the waterfall on the right at close range. Vegetation rushes past both sides of the frame with physically correct directional motion blur, while the hands, housing and central section of tape remain mostly sharp and stable. Use a powerful smooth speed ramp, immersive but not shaky.
5.5–8.5 seconds:
Without cutting, the dense vegetation opens naturally and the camera bursts out over a wide turquoise tropical river. Fly only slightly above the water while continuing to follow the same uninterrupted yellow tape stretched across its surface. Realistic water spray, sunlight reflections and surrounding jungle cliffs rush beneath and beside the viewer. A modern city skyline gradually becomes visible ahead through the humid atmosphere. The transition happens through forward travel and natural environmental reveal, never through visible image morphing.
8.5–12 seconds:
The camera reaches the city waterfront and continues at high speed into a beautiful modern business district. Glide between realistic glass office towers and along a clear urban corridor, always following the same centered yellow tape. Reflections slide naturally across the façades. Background architecture and side edges show strong speed blur, but the hands, tape housing and yellow guide line remain coherent and readable. The headquarters becomes visible directly ahead and grows naturally larger as the camera approaches.
12–14 seconds:
Enter the open plaza in front of the headquarters and smoothly decelerate. The motion blur gradually disappears. Follow the yellow tape precisely toward the central glass entrance without changing direction. Stop only a few metres from the doors, matching the camera height, hand placement, tape length, scale, lighting and composition of @Image2.
14–15 seconds:
Settle perfectly into the exact final frame @Image2 and hold it stable. The hands make a tiny natural settling movement and then remain still. The yellow tape ends exactly at the entrance. Preserve the complete emblem and wordmark above the doors with the exact original design, gradient colors, proportions, capitalization and spelling. The logo must be sharp, fully readable and completely stable.
Хотите сгенерировать ролик со своей локацией? Закиньте промпты в новую Gemini 3.6 Flash. Она отлично понимает контекст и преобразует тексты
Правда, на официальном сайте с ней могут быть проблемы из-за сложной проверки VPN. Попробуйте свеженькую модель на SpeShu.AI без лишних трудностей.
По подписке ещё выгоднее: за 1 690 рублей вы получите сопоставимые вариации не только Gemini, но Claude c ChatGPT (вместо 6 000 рублей).
Ещё выгоднее с промокодом
Правда, на официальном сайте с ней могут быть проблемы из-за сложной проверки VPN. Попробуйте свеженькую модель на SpeShu.AI без лишних трудностей.
По подписке ещё выгоднее: за 1 690 рублей вы получите сопоставимые вариации не только Gemini, но Claude c ChatGPT (вместо 6 000 рублей).
Ещё выгоднее с промокодом
TG15.👍4🤯2💯2
Итальянки позавидуют вашей красоте 💃
В Фотостудии SpeShu.AI новая фотосессия — летняя dolce vita!
Попробуйте бесплатно, у вас на балансе 18 токенов 🎁
В Фотостудии SpeShu.AI новая фотосессия — летняя dolce vita!
Попробуйте бесплатно, у вас на балансе 18 токенов 🎁
👍3🤯3💯2
Media is too big
VIEW IN TELEGRAM
Как создать рекламу на 500 000 рублей, не имея ни копейки
Пресет и один промпт — вот что нужно, чтобы создать стильное рекламное видео в 2026 году. Генерируем товар и локацию в ChatGPT Image 2.0, затем анимируем в Seedance 2.0:
Пресет и один промпт — вот что нужно, чтобы создать стильное рекламное видео в 2026 году. Генерируем товар и локацию в ChatGPT Image 2.0, затем анимируем в Seedance 2.0:
Create an exactly 15-second ultra-photorealistic vertical 9:16 luxury automotive commercial at 24 fps.
Use @image1, @image2 and @image3 as exact chronological storyboard references. They show the same single customized black-and-gold Rolls-Royce Spectre. Preserve the exact vehicle identity, proportions, glossy obsidian-black paint, gold Pantheon grille, gold Spirit of Ecstasy, forged carbon with irregular gold flecks, exact gold disc wheels and every visible design detail.
The images are cinematic keyframes, not slides. Generate real camera movement inside every shot and use sharp commercial hard cuts. Never morph, dissolve or blend the reference images together.
0–4.5 seconds:
Begin directly from @image1. Extreme macro of the gold Spirit of Ecstasy above the forged-carbon hood. The ornament and vehicle remain stationary. The camera performs a very slow push-in combined with a subtle left-to-right micro-orbit. A narrow soft reflection travels naturally across the gold and carbon surfaces. Shallow optical depth of field, realistic focus breathing, restrained sensor grain and physically accurate highlight roll-off.
4.5–9.5 seconds:
Hard cut on a deep musical beat to @image2. Low symmetrical frontal view of the exact grille and split headlights. The camera slowly moves from a slight front-left position into perfect center alignment while gently pulling backward. The headlights smoothly rise to a restrained steady white intensity. Reflections travel coherently across the long hood, gold grille and polished black floor.
9.5–15 seconds:
Hard match cut from a gold grille reflection to the exact front-left wheel in @image3. Ground-level close-up. The wheel and car remain stationary while the camera performs a slow controlled arc around the wheel and travels slightly along the fender toward the forged-carbon side skirt. A warm band of light glides across the machined gold wheel face, revealing realistic metal texture, tire rubber and carbon detail. During the final second, gently push closer until the gold wheel fills most of the frame, creating a strong visual endpoint for the next edit.
High-budget real automotive cinematography, dark charcoal-stone gallery, restrained black-and-gold palette, physically accurate materials and reflections, realistic exposure, subtle natural motion blur, deep detailed shadows, no CGI appearance.
Sound: minimal dark luxury soundtrack, restrained bass impact at every cut, subtle gallery ambience, no narration.
Strict constraints: exactly three shots in the stated order; no slideshow, crossfades, morphing or frozen images; do not redesign the vehicle; no wheel rotation, vehicle movement, warped geometry, duplicated ornament, altered grille, people, extra vehicles, generated captions or watermark. Create an exactly 10-second ultra-photorealistic vertical 9:16 continuation of the same luxury automotive commercial at 24 fps.
Use @image1 as the mandatory opening side-profile composition and @image2 as the final hero-shot reference. Preserve the exact same customized black-and-gold Rolls-Royce Spectre, including its body proportions, glossy obsidian-black paint, forged carbon with gold flecks, gold trim, exact gold disc wheels and every visible design detail.
IMPORTANT OPENING: Begin instantly from the wide side-profile angle in @image1. This must sharply contrast with the wheel close-up that ended the previous video. No fade-in, black screen or repeated wheel macro. The change from an extreme detail to the complete vehicle must feel like an intentional cut on a musical beat.
🤩3💯3🤯1
Вторая часть промпта для рекламной перезентации:
0–4.5 seconds:
Begin directly from @image1 with the complete left-side profile visible. The stationary car remains fully inside the vertical frame. The camera performs a perfectly smooth parallel tracking movement from the rear section toward the front. A long warm reflection travels across the lower doors while cool overhead light defines the roofline and shoulder. Both wheels remain perfectly circular and unchanged. Realistic tire contact, contact shadow and subdued floor reflection.
4.5–10 seconds:
Hard cut on a strong bass impact to the low front-left three-quarter hero composition in @image2. Keep the complete vehicle visible. The camera slowly dollies backward while making a restrained eight-degree arc around the front, creating natural parallax between the grille, front wheel and long side body. The headlights glow at a steady realistic intensity. A faint warm backlight traces the roof and shoulder line. Subtle haze remains only behind the vehicle.
During the final second, the camera gently settles and becomes almost stationary, leaving clean negative space above the car for a logo or text to be added during editing. Do not generate any text.
High-budget real automotive cinematography, dark monumental gallery, black-and-gold luxury palette, physically accurate optics, materials, illumination and reflections, restrained sensor grain, no CGI appearance.
Strict constraints: exactly two shots; no fade-in, slideshow, dissolve or image morphing; the vehicle remains stationary; no changing wheels, body deformation, altered grille, extra vehicles, people, generated captions or watermark.
✍3💯2👨💻2
Возникла проблема с промптом или хотите сильно адаптировать его под свой товар?
Используйте SpeShu Claude. Эта нейросеть практически не уступает западной модели. При этом стоит в разы дешевле.
Второй важный плюс: ваш аккаунт не заблокируют. Вам не нужно скрывать российскую почту и IP, чтобы защитить свои данные от бана Антропиков.
Попробовать SpeShu Claude + SpeShu Gemini и ChatGPT можно по подписке. Откройте личный кабинет на SpeShu.AI и выберите оптимальный для вас тариф.
Используйте SpeShu Claude. Эта нейросеть практически не уступает западной модели. При этом стоит в разы дешевле.
Второй важный плюс: ваш аккаунт не заблокируют. Вам не нужно скрывать российскую почту и IP, чтобы защитить свои данные от бана Антропиков.
Попробовать SpeShu Claude + SpeShu Gemini и ChatGPT можно по подписке. Откройте личный кабинет на SpeShu.AI и выберите оптимальный для вас тариф.
👍3👨💻3🤯2
Media is too big
VIEW IN TELEGRAM
Аниме в Seedance 2.0
Генерируем видео на 30 секунд двумя промптами. Ловите первый:
Генерируем видео на 30 секунд двумя промптами. Ловите первый:
[Reference Lock] @image1 defines the ONLY visual identity for the courier girl — her face, hairstyle, outfit, goggles, satchel, and her wooden propeller-glider — matching @image1 exactly in every shot, no identity drift. @image2 (location reference) defines the hillside village, windmills, cliffs, and waterfall setting. Match the starting frame exactly as the opening shot and camera position.
Widescreen 16:9, hand-drawn Studio Ghibli / Hayao Miyazaki animation style, painterly 2D cel-shaded rendering with soft watercolor background art, warm golden-hour cinematic lighting, 15 seconds, 4 DISTINCTLY DIFFERENT camera shots with clean hard cuts between them — each shot must use a completely different camera angle, distance, and height, no two shots may share the same framing. Dynamic flight choreography, atmospheric and windswept throughout, no dialogue, no text, no subtitles.
Ultra high quality, stable consistent character and glider geometry throughout every shot, consistent golden-hour lighting continuity, no morphing, no flickering, temporally stable painterly textures.
SHOT 1 (0-4s) — WIDE TRACKING SHOT flying alongside her at eye level: she soars triumphantly above the village holding up the letter, wind streaming through her hair, when a sudden strong gust curls around her hand. In a flash the letter is torn free, tumbling away into the open air. Her eyes go wide with alarm, and she banks the glider hard to give chase.
CUT TO SHOT 2 (4-8s) — LOW UPWARD HERO ANGLE from village rooftop level: the letter tumbles and flutters down between two tall windmills, their great wooden blades slowly turning. She dives the glider steeply between the windmill blades in a tight, thrilling near-miss maneuver, skirts fluttering, one blade sweeping dramatically past just behind her.
CUT TO SHOT 3 (8-12s) — DYNAMIC SIDE TRACKING SHOT racing alongside her at full speed: she swoops low over cascading rooftop laundry lines, white sheets billowing and snapping in her wake, chasing the letter as it weaves erratically through the wind above the cobblestone streets, villagers' washing rippling like waves below her.
CUT TO SHOT 4 (12-15s) — WIDE CINEMATIC SHOT pulling back to reveal scale: the letter suddenly catches an updraft and rockets upward toward a cliff-edge waterfall glowing gold in the sunset light, disappearing into a swirl of loose flower petals and dust caught in the wind. She pulls the glider up hard to follow, small against the vast golden sky and cliffside. Freeze on this wide triumphant-but-tense final frame, the letter just visible vanishing into the petal-swirl ahead of her — end of Part 1.
Sound design: rushing wind, fluttering fabric, a soft gasp, creaking windmill blades, flapping laundry sheets, the whoosh of the glider banking and diving. Sweeping, whimsical Ghibli-style orchestral score with flutes and strings, building adventurous energy throughout, no dialogue, no vocals.
Negative prompt: no text, no subtitles, no watermark, no logos, no character identity drift, no changes to her face, outfit, or glider design, no morphing geometry, no flickering textures, no extra or malformed fingers, no duplicated objects, no static or flat camera work, no modern anime style, no digital flat-shading — must maintain painterly hand-drawn Ghibli texture throughout.
Второй промпт для аниме в Seedance 2.0:
Хотите сгенерировать видео со своим персонажем, по своему сюжету? Загрузите эти промпты в SpeShu Claude — эта российская вариация топового ИИ быстро схватит паттерны, стиль, настроение и механику промптинга. Вам останется лишь описать свою идею, и она отправит вам готовые запросы для ChatGPT Image 2.0 и того же Seedance 2.0.
Попробуйте SpeShu Claude, а также SpeShu Gemini и ChatGPT в 3 раза дешевле, чем по зарубежным подпискам. Для этого зарегистрируйтесь на SpeShu.AI, перейдите в личный кабинет, а затем в раздел подписок.
[Reference Lock]
Image 1 (character reference sheet) defines the girl's face and identity.
Image 2 (location reference) defines the hillside village, cliffs, and waterfall setting.
Image 3 (outfit-and-glider reference photo) defines her EXACT outfit — mustard-yellow vest over cream blouse, brown suspender shorts, knee socks, boots, satchel, braid, goggles pushed up on forehead — and the EXACT wooden hand-glider design with patchwork fabric wing panels. The opening shot of this generation must visually match Image 3's framing closely: her and the glider clearly, fully visible in frame, not obscured or cropped.
This continues directly from the chase established in Part 1: she has just followed a wind-snatched letter into a swirl of petals near a cliff-edge waterfall glowing gold in the sunset light.
Widescreen 16:9, hand-drawn Studio Ghibli / Hayao Miyazaki animation style, painterly 2D cel-shaded rendering with soft watercolor background art, warm golden-hour cinematic lighting, 14-15 seconds, 4 DISTINCTLY DIFFERENT camera shots with clean hard cuts between them — each shot must use a completely different camera angle, distance, and height. Dynamic flight choreography, atmospheric and windswept throughout, no dialogue, no text, no subtitles.
Ultra high quality, stable consistent character and glider geometry throughout every shot, matching Image 3's design exactly, consistent golden-hour lighting continuity, no morphing, no flickering, temporally stable painterly textures.
SHOT 1 (15-19s) — WIDE CLEAR ESTABLISHING SHOT, camera positioned to fully frame both the girl and her glider exactly as in Image 3, no cropping, no obstruction: she banks the glider sharply, fully visible astride the wooden frame, reaching toward the letter tumbling just ahead of her through drifting petals and mist near the waterfall's edge, sunset light glinting off the wing's patchwork fabric.
CUT TO SHOT 2 (19-22s) — CLOSE TRACKING SHOT flying just behind her shoulder: she banks through the waterfall's mist, droplets catching golden light like scattered jewels around her, the letter tumbling just ahead through the spray, always a hair's breadth out of reach.
CUT TO SHOT 3 (22-25s) — WIDE DRAMATIC SHOT from below looking up through the mist, glider and girl both clearly silhouetted and readable: she makes one final desperate lunge, snatching the letter cleanly from the air mid-dive, clutching it triumphantly to her chest, a bright relieved smile breaking across her face.
CUT TO SHOT 4 (25-30s) — WIDE EPIC PULL-BACK SHOT rising and receding, glider and rider kept clearly visible and well-lit throughout, never silhouetted into unreadable shadow: she tucks the letter safely into her satchel and banks the glider gracefully toward the setting sun, soaring out over the sparkling sea beyond the cliffs, the village and windmills glowing gold behind her. Camera continues pulling back into a vast, breathtaking wide shot, keeping her and the glider as a clearly readable focal point against the enormous glowing sunset sky. Natural, warm, epic ending.
Sound design: rushing waterfall spray, a triumphant gasp, delighted laughter, distant village bells, wind easing into a gentle breeze. Sweeping Ghibli-style orchestral score swelling to an emotional triumphant peak at the catch, then resolving into a warm, spacious final theme with flutes and strings, no dialogue, no vocals.
Negative prompt: no text, no subtitles, no watermark, no logos, no character identity drift, no changes to her face, outfit, or glider design, no morphing geometry, no flickering textures, no extra or malformed fingers, no duplicated objects, no static or flat camera work, no modern anime style, no digital flat-shading, no obscured or poorly visible character/glider in any shot, no silhouette-only framing that hides outfit or glider details, no aviator helmet, no goggles worn over the eyes, no dress or skirt.
Хотите сгенерировать видео со своим персонажем, по своему сюжету? Загрузите эти промпты в SpeShu Claude — эта российская вариация топового ИИ быстро схватит паттерны, стиль, настроение и механику промптинга. Вам останется лишь описать свою идею, и она отправит вам готовые запросы для ChatGPT Image 2.0 и того же Seedance 2.0.
Попробуйте SpeShu Claude, а также SpeShu Gemini и ChatGPT в 3 раза дешевле, чем по зарубежным подпискам. Для этого зарегистрируйтесь на SpeShu.AI, перейдите в личный кабинет, а затем в раздел подписок.
Claude Opus 5 уже здесь 🔥
Изучили, что нового придумали Антропик — почитайте в новой статье: https://habr.com/ru/companies/tsnis/articles/1062830/
Уже на SpeShu.AI без VPN и зарубежной подписки. Тестируйте с ещё большей выгодой — примените промокод
Изучили, что нового придумали Антропик — почитайте в новой статье: https://habr.com/ru/companies/tsnis/articles/1062830/
Уже на SpeShu.AI без VPN и зарубежной подписки. Тестируйте с ещё большей выгодой — примените промокод
TG15.👍4⚡3🤯2💯1