библиотека
Готовые промпты из статей блога. Выберите тему, скопируйте промпт, замените данные в скобках на свои. Новые статьи добавляют сюда свои промпты сами.
Claude сам проходит все шаги гайда и сохраняет готовый ролик MP4 в папку.
Прочитай гайд https://digitolga.blog/opus-motion-design/ и сделай всё по шагам. Сначала установи всё, что нужно (шаг 1). Потом сделай ролик: [о чём ролик], [длительность в секундах], [формат: вертикальный 9:16 или горизонтальный 16:9]. Работай в HyperFrames. Перед полным рендером покажи мне 5 стоп-кадров и оцени их по критериям из шага 7.
Ролик-визитка на 15 секунд. Быстрая проверка, что всё настроено и Opus умеет анимировать.
make a dynamic 15-second motion graphics video that shows what an incredible motion designer you are, like it's your showreel for a résumé. go all out.
Ролик на 10 секунд без рамок и подписей в углах, по которым сразу видно ИИ.
make a dynamic 10-second motion graphics video that introduces who you are as Opus 5.5. Avoid frames and text in the corners, the usual giveaways of AI-made video.
Вертикальный ролик на 45 секунд: история вашей темы от начала до сегодня. Раскадровку Opus придумает сам.
use your showreel energy, but tell a story: the history of [TOPIC] from [START] to today, surprise me with the storyboard. 45 seconds, vertical 9:16.
Бриф из восьми строк, после которого Opus делает ваш ролик, а не шаблонный.
GOAL: [после ролика зритель понимает, что кофе можно заказать за два нажатия] OUTPUT: HyperFrames, размер [1080 x 1920] пикселей ([вертикальный]), [30] кадров в секунду TIME: длина [15] секунд, [без петли] LOOK: фон [#0D0A08], текст [#F5F0E8], акцентный цвет [#F5D347], шрифт [Oswald 700, 96 px] MOTION: появления с плавным торможением [cubic-bezier(0.16, 1, 0.3, 1)], пружины с затуханием [0.8], каждое движение длится [0,4] секунды SHOTS: [0–3 с: логотип собирается из букв] / [3–8 с: телефон выезжает, курсор нажимает «Заказать»] / [8–15 с: карточка с ценой и логотипом] BANNED: [заголовок по центру на градиенте, свечение, частицы, подписи в углах] QA: до полной сборки покажи стоп-кадры на секундах [0, 3, 8, 12, 15] и все кадры на одной картинке
Файл правил для папки проекта вместе с техническими правилами: ролик не мерцает и не дрожит, а Claude сам проверяет кадры.
# Motion studio rules ## Render contract - Build every film in HyperFrames and render it with HyperFrames, H.264 yuv420p. - Only the libraries the brief names, pinned to a version. No network requests except fonts and those libraries. - Every frame is a pure function of time t, in seconds: `window.seek(t)` paints frame t. - Animate with CSS keyframes, the Web Animations API, SVG, Canvas 2D or WebGL. - No CSS transitions, no Date.now, no setTimeout or setInterval, no requestAnimationFrame in render mode, no state carried between frames. Seeded noise only (mulberry32), never Math.random. - Loops: render frames 0 to N-1 only, and frame N-1 must flow into frame 0. ## Motion - Arrive fast, land soft: each frame covers 12–19% of the remaining distance. Never constant speed, never a dead stop. - Nothing ever freezes: holds keep a slow push-in (~0.25% per frame). - Every move lasts at least 0.3s; big moves 0.5–0.75s. - Stagger grouped elements 2–4 frames apart. - Motion blur follows direction and speed, fades as things settle; fast moves get real in-between frames blended, never blended across a cut. - Text stays still at least 8 frames before it moves. - One focal point per frame. - Cuts land on the beat or 2 frames before it; morphs start ~4 frames early. ## Look - Banned defaults: centered title on gradient, everything fading in, corner labels and frame borders, glow on UI chrome, generic particle bursts. - One display face, one UI face. One accent color unless the brief says otherwise. - Every 2 to 4 seconds something new must happen on screen. ## Sound - Score and SFX are synthesized in code unless a track is supplied. - Place hits on the measured beat grid (beats.json). Loudness -14 LUFS. ## Loop before you show me anything 1. Render one frame per beat as a contact sheet and LOOK at it. 2. Score it 1-10 on: hook in first 2s, readability at phone size, motion quality, variety, brand accuracy, sound sync. 3. Fix the 3 worst problems. Repeat until every score is 8+. 4. Only then do the full render.
Покадровый разбор чужого ролика в цифрах: длительности, позиции, цвета, удары звука.
Before building anything, reverse-engineer [./ref.mp4] frame by frame. 1. Extract every frame. Find every cut from frame-to-frame difference spikes, then confirm each one visually. 2. For every shot, measure with code (never eyeball): start/end frame, each element's position and size per frame, how many frames each move takes, overshoot, camera zoom/pan per frame, blur timing and direction, exact colours sampled from pixels, font sizes, and every sound hit's timestamp. 3. Save it as SPEC.md with the per-frame numbers for every big move. The frames are the truth. If your description and the frames disagree, the frames win. Show me the shot list before you build.
Два списка: что взять из референса один в один, а что заменить на ваш бренд.
Before building, split the reference into two lists: KEEP (identical): timing, cuts, camera, easing, positions, blur, transitions. CHANGE (mine): brand, logo, colours, fonts, copy, product screens, faces, platform UI. Then write my visual rules once, in one place (fonts, palette, spacing, one accent colour) and every shot must use them. BANNED: random gradients, everything fading in, fake product UI, particle bursts, glow on text, more than one thing fighting for attention. Ask me about anything ambiguous before you build.
Пять ключевых кадров на проверку до полного рендера. Экономит время на переделках.
Before the full render, render 5 stills: the opening, the main composition, the product shot, the middle of the fastest transition, and the end card. Look at each one yourself, zoomed in, and check: hierarchy, spacing, typography, colours, product accuracy, and anything that shouldn't be there (stray shapes, cut-off text, overlaps, leftover effects you copied without knowing why). Fix everything, re-render the stills, then show me. Only start the full render after I approve.
Оценка ролика по семи критериям и исправление трёх худших мест, пока всё не дойдёт до 8 из 10.
Render a contact sheet (2 frames per second), a strip of 12 frames around each fast move, and a phone-size sheet at 360 px wide. Look at them properly. Be a harsh motion director, not a proud author. Score 1-10: hook in the first 2s, readability at phone size, motion quality, variety (something new every 2-4s), composition, data accuracy, sound sync. List the 3 biggest problems with timestamps. Fix them. Re-render only the affected seconds. Repeat until every score is 8+.
Строгая рецензия: каждое расхождение с референсом с номером кадра и готовым исправлением.
You're a harsh motion director and you didn't build this. Default: reject.
Put the reference and my render side by side, frame-locked, and compare them.
For every mismatch give: shot + frame number, what's wrong, why it looks
worse, and the exact fix ("card lands 6 frames early, delay to frame 44").
Also measure frame-to-frame change and flag:
- any fast move followed by a dead stop (that's what "choppy" is)
- text moving before it's readable
- sound landing off its visual hit
Prioritize the 5 fixes that improve the video most. Nothing ships below [8]/10.
Оценка поста до публикации: цепляет ли первая строка и захотят ли его сохранить.
{
"hook_strength": {
"type": "score",
"instructions": "How strong is the first line as a scroll-stopping hook?",
"criteria": [
"Generic opener, nothing to stop for",
"Weak: needs the rest of the post to make sense",
"Fine: clear, not surprising",
"Strong: a concrete promise, number or twist",
"Impossible to skip"
]
},
"save_likelihood": {
"type": "score",
"instructions": "How likely is a reader to save this post for later?",
"criteria": [
"Very unlikely",
"Unlikely",
"Possible",
"Likely",
"Very likely"
]
},
"specificity": {
"type": "score",
"instructions": "How specific is this post: concrete numbers, names or examples?",
"criteria": [
"Only general statements",
"One vague example",
"Some specifics, mostly general",
"Mostly concrete",
"Every paragraph has a fact, number or example"
]
},
"single_point": {
"type": "noul",
"instructions": "The post makes one clear point."
},
"filler": {
"type": "noul",
"instructions": "The post contains corporate filler or empty generalisations."
},
"ai_text": {
"type": "noul",
"instructions": "The post reads as generic AI-generated text."
},
"intent": {
"type": "choice",
"instructions": "What is the main intent of this post?",
"criteria": {
"teach": "Explains how to do something",
"sell": "Promotes a service or product",
"entertain": "Amuses or tells a story",
"announce": "Shares news",
"opinion": "States a personal view"
}
},
"audience": {
"type": "choice",
"instructions": "Who is this post written for?",
"criteria": {
"beginner_creator": "Someone starting with AI content",
"pro_creator": "An experienced creator",
"brand_manager": "A marketing lead at a brand",
"potential_client": "Someone who might hire the author",
"nobody": "No specific reader"
}
}
}
Отбор сгенерированных кадров: какие годятся для рекламы, а какие с браком.
{
"usable": {
"type": "score",
"instructions": "How professional and commercially usable does this shot sound?",
"criteria": [
"Amateur, unusable",
"Weak",
"Acceptable for social",
"Professional",
"Campaign quality"
]
},
"anatomy": {
"type": "noul",
"instructions": "The description mentions distorted hands, faces or anatomy."
},
"garbled_text": {
"type": "noul",
"instructions": "The description mentions unreadable or garbled text in frame."
},
"premium_light": {
"type": "noul",
"instructions": "The lighting described matches a premium editorial look."
},
"mood": {
"type": "choice",
"instructions": "What is the dominant mood of this shot?",
"criteria": {
"luxury_calm": "Quiet, expensive, unhurried",
"raw_documentary": "Real, unpolished, candid",
"playful_bright": "Light, colourful, fun",
"dark_cinematic": "Low key, dramatic",
"flat_boring": "No mood at all"
}
},
"platform": {
"type": "choice",
"instructions": "Which platform does this shot fit best?",
"criteria": {
"instagram_reels": "Vertical video, fast hook",
"instagram_feed": "Still image for the grid",
"client_presentation": "A deck for a brand",
"website_hero": "A large hero image",
"nowhere": "Not usable anywhere"
}
}
}
Карта приёмов конкурентов: какие крючки и обещания используют чужие объявления.
{
"hook_type": {
"type": "choice",
"instructions": "What hook type does this ad use in the first three seconds?",
"criteria": {
"question": "Opens with a question",
"bold_claim": "Opens with a strong statement",
"problem": "Shows a pain point",
"before_after": "Contrast of before and after",
"demo": "Product in use",
"face": "A celebrity or a person talking to camera",
"none": "No visible hook"
}
},
"offer": {
"type": "choice",
"instructions": "What is the core offer?",
"criteria": {
"discount": "Price cut or promo",
"free_trial": "Try before you buy",
"new_product": "A launch",
"social_proof": "Reviews, numbers, testimonials",
"education": "Teaches something",
"brand_image": "Image only, no offer"
}
},
"production": {
"type": "choice",
"instructions": "What production format is this ad?",
"criteria": {
"ugc_selfie": "Phone selfie style",
"studio_product": "Studio product shot",
"motion_graphics": "Animation or motion graphics",
"ai_generated": "Visibly AI-generated",
"lifestyle_film": "Filmed lifestyle scene"
}
},
"cta_clarity": {
"type": "score",
"instructions": "How clear is the call to action?",
"criteria": [
"No call to action",
"Vague",
"Present but weak",
"Clear",
"Impossible to miss"
]
},
"emotional": {
"type": "score",
"instructions": "How emotional is this ad rather than rational?",
"criteria": [
"Purely rational",
"Mostly rational",
"Balanced",
"Mostly emotional",
"Purely emotional"
]
},
"segment": {
"type": "noul",
"instructions": "The ad names a specific audience segment."
}
}
Сортировка комментариев и сообщений: настоящие вопросы отдельно, спам и похвала отдельно.
{
"kind": {
"type": "choice",
"instructions": "What kind of message is this?",
"criteria": {
"real_question": "A genuine question to the author",
"praise": "A compliment",
"criticism": "Substantive criticism",
"spam": "Spam or self-promotion",
"aggression": "Hostile or abusive",
"unclear": "Cannot tell"
}
},
"deserves_reply": {
"type": "noul",
"instructions": "This message deserves a personal reply from the author."
},
"business": {
"type": "noul",
"instructions": "This message contains a business enquiry."
},
"aggression": {
"type": "score",
"instructions": "How aggressive is the tone?",
"criteria": [
"Friendly",
"Neutral",
"Irritated",
"Hostile",
"Abusive"
]
}
}
Очередь заявок: сначала те, где есть бюджет и задача из вашей зоны.
{
"budget": {
"type": "noul",
"instructions": "The enquiry mentions a budget or budget range."
},
"deadline": {
"type": "noul",
"instructions": "The enquiry mentions a deadline."
},
"in_scope": {
"type": "noul",
"instructions": "The requested work is AI video, AI image or AI animation production."
},
"free_work": {
"type": "noul",
"instructions": "The sender asks for free work, a test task or a big discount."
},
"clarity": {
"type": "score",
"instructions": "How clearly is the task described?",
"criteria": [
"No task described",
"Vague",
"Some details",
"Clear",
"Fully specified: format, length, deadline, references"
]
},
"paid_likelihood": {
"type": "score",
"instructions": "How likely is this enquiry to become a paid project?",
"criteria": [
"Very unlikely",
"Unlikely",
"Possible",
"Likely",
"Very likely"
]
},
"client_type": {
"type": "choice",
"instructions": "What type of client is writing?",
"criteria": {
"brand": "A brand or company",
"agency": "An agency or studio",
"private": "A private person",
"creator": "Another creator",
"recruiter": "A recruiter",
"unclear": "Cannot tell"
}
}
}
Самые сильные фрагменты длинного видео для коротких клипов.
{
"moment_strength": {
"type": "score",
"instructions": "How strong is this fragment as a standalone short clip?",
"criteria": [
"Nothing happens",
"Weak",
"Average",
"Strong",
"A clip people would share"
]
},
"surprise": {
"type": "score",
"instructions": "How surprising or counterintuitive is the statement here?",
"criteria": [
"Obvious",
"Slightly unexpected",
"Somewhat surprising",
"Surprising",
"Counterintuitive, makes you stop"
]
},
"standalone": {
"type": "noul",
"instructions": "This fragment makes sense without the rest of the video."
},
"concrete": {
"type": "noul",
"instructions": "This fragment contains a concrete number, name or example."
},
"best_use": {
"type": "choice",
"instructions": "What is the best use of this fragment?",
"criteria": {
"opening_hook": "Opening of the short",
"standalone_short": "A short on its own",
"b_roll": "Supporting footage",
"cut": "Cut it"
}
}
}
Проверка текста на ИИ-штампы: пустые вступления, повторы, «не просто X, а Y».
{
"not_just": {
"type": "noul",
"instructions": "The text uses the pattern 'not just X, but Y'."
},
"empty_opener": {
"type": "noul",
"instructions": "The text opens with a generic scene-setting sentence that carries no information."
},
"repeats": {
"type": "noul",
"instructions": "The text repeats the same idea in different words."
},
"cliches": {
"type": "noul",
"instructions": "The text uses corporate or marketing cliches."
},
"concrete_paragraphs": {
"type": "noul",
"instructions": "Every paragraph contains at least one concrete fact, number or example."
},
"human": {
"type": "score",
"instructions": "How human and specific does this text sound?",
"criteria": [
"Template text",
"Mostly generic",
"Mixed",
"Mostly human",
"A specific person is clearly talking"
]
}
}
Ничего не нашлось. Попробуйте другое слово или снимите фильтр.
Больше про ИИ в моём Telegram
Подписаться →