TGViewer
AI Роман | Про нейросети AI Роман | Про нейросети @generated_roman · 6.86K subscribers
Post #925 1K
🏗 Повторяем самый хайповый тренд в инсте*

Мне не особо нравится, что лента в моменте забита подобными видосами, но это тренды, и вы сами их просите, поэтому рассказываю ⤵️

Как повторить:
1⃣ Заходим в GPT Image или Nano Banana, прикрепляем своё фото в качестве референса и вставляем промпт:
Use the uploaded photo as the identity reference for the person: preserve the overall identity, silhouette, and recognizability from the reference, but do not redesign the model. Create a realistic cinematic photo in 9:16 vertical format.

Show
one single person standing on the very top of a skyscraper antenna / mast, high above New York City. The composition is dramatic and vertigo-inducing: the person is positioned near the top center-left of the frame, balancing on the small metal structure at the summit, while holding a large black flag that extends diagonally across the frame. The background shows a hazy panoramic Manhattan skyline far below, with the Empire State Building clearly visible on the right side and the river plus dense city blocks fading into atmospheric distance.

The person should wear the same style of outfit as in the reference image:
all-black outfit, including a black sleeveless fitted top, black pants, black shoes, and a black face covering / balaclava or mask, creating a bold minimalist urban look. Keep the outfit clean, tactical, and striking.

The flag should be
large, black, and wind-blown, attached to a pole and occupying a big portion of the middle-right of the image. Leave a clearly visible area for text on the flag. Render the flag text as [ТЕКСТ НА ФЛАГЕ] in a bold uppercase condensed sans-serif font, large white letters, centered and highly legible.

The antenna structure should include realistic industrial details: metal ladder sections, safety rails, a narrow platform, and a small red warning beacon light near the top. The overall framing should feel like a high-end extreme urban photography shot captured with a long lens: compressed city background, realistic scale, and intense sense of height.

Lighting should be
soft daylight / overcast haze, with natural muted colors, subtle contrast, and realistic shadows. Style: ultra-realistic photography, cinematic tension, documentary/editorial feel, high detail, slight atmospheric haze, no fantasy look.

Negative prompt: two people, extra people, different outfit colors, cartoon, CGI, surreal fantasy, low detail city, distorted anatomy, extra limbs, unreadable flag, messy text, low resolution, watermark, logo, text outside the flag.

2⃣ Переходим на платформу, где есть SeeDance 2.0 (здесь рассказывал, как получить доступ к нейронке)
3⃣ Вставляем получившееся фото и промпт (оставил его в комментариях)

🔥 – нужно больше трендовых инструкций

*запрещена в рф
  • 🔥 15
  • ❤ 2
  • ❤‍🔥 2
  • 🤣 2
More from @generated_roman
  1. Aug 24, 2026Ты едешь или нет? Нет. Читал новости, вспомнил мем, не смог удержаться, чтобы не сделать э…
  2. Aug 19, 2026🤯 Если нет разницы – зачем платить больше? Затестил SeeDance 2.5 и Gemini Omni Flash на с…
  3. Aug 10, 2026🤪 Поясняю за цены на SeeDance 2.5 🍀 Dreamina AI - официальный дилер, 30 секунд в разреше…
  4. Aug 5, 2026🎙 Как бесплатно перевести любое аудио в текст? Рубрика #эксперименты. Собрал для вас Tele…
  5. Aug 4, 2026📌 ПОДБОРКА ПОЛЕЗНЫХ ПОСТОВ НА КАНАЛЕ Собрал в одном месте все актуальные посты, которые с…
  6. Jul 30, 2026Как вам такой способ? 😏
Threads Profile ViewerView any public Threads profile without an account.Open ThreadLook →Writing with AI? Make it sound human.Metric37 rewrites AI drafts so they read naturally. Free AI detector, 1,500 words free.Try Metric37 →