Trong lĩnh vực sáng tạo video bằng AI, việc xây dựng một quy trình (workflow) có logic rõ ràng quan trọng hơn rất nhiều so với việc viết prompt rời rạc. Bộ prompt dưới đây giúp bạn tạo ra một video thời trang hoàn chỉnh theo concept:
👉 “Ngủ dậy → Nhận đồ → Xem đồ → Thay đồ → Catwalk & tạo dáng thời trang”
Đây là dạng nội dung cực kỳ phù hợp để:
- Làm video TikTok/Reels
- Quảng bá sản phẩm thời trang
- Tạo content lifestyle tự nhiên, không bị “quảng cáo lộ liễu”
1. Tổng quan quy trình video
Quy trình được chia thành 5 giai đoạn chính:
Giai đoạn 1: Ngủ dậy (Lifestyle – Tự nhiên)
- Nhân vật vừa thức dậy
- Tạo cảm giác chân thật, gần gũi
- Outfit: đồ ngủ (pajamas)
Giai đoạn 2: Nhận & mở hộp sản phẩm
- Nhân vật cầm hộp hàng
- Mở hộp nhẹ nhàng
- Tương tác với sản phẩm bên trong
Giai đoạn 3: Xem & cảm nhận trang phục
- Nhân vật ngồi trên giường
- Cầm và quan sát quần áo
- Chuyển trạng thái từ “sleepy” → “quan tâm sản phẩm”
Giai đoạn 4: Thay đồ (Chuyển outfit)
- Chuyển từ đồ ngủ → outfit chính
- Đây là điểm “hook” quan trọng nhất của video
Giai đoạn 5: Trình diễn thời trang
- Walk nhẹ
- Nhiều góc quay: high angle, low angle, back shot, close-up
- Tạo cảm giác chuyên nghiệp như fashion shoot
2. Điểm mạnh của bộ prompt này
Bộ prompt bạn đang dùng có một số ưu điểm cực kỳ quan trọng:
✔ Giữ nguyên identity nhân vật
- “Use the same face identity from reference image #1”
- Tránh lỗi AI bị biến mặt giữa các frame
✔ Không thay đổi outfit sai logic
- Luôn ràng buộc:
- Outfit từ image #2
- Pajamas từ image #3
- Tránh lỗi “AI tự đổi đồ”
✔ Kiểm soát chuyển động tự nhiên
- Nhấn mạnh:
- “slow, natural movement”
- “subtle motion”
- Giúp video không bị robotic
✔ Ràng buộc kỹ thuật chặt chẽ
- 4K / 60FPS
- No flicker / no morphing
- Accurate anatomy
👉 Đây là yếu tố khiến video trông “real” thay vì “AI giả”
3. Breakdown chi tiết từng prompt
3.1. Cảnh ngủ dậy
| Hình ảnh | Prompt ảnh |
 |
Messy hair, no makeup, wearing pajamas, eyes closed, cute sleepy expression, stretching arms, soft bedroom lighting. Keep the face from the reference image exactly the same, but completely change the pose. |
| Prompt video |
| Use the same face identity from reference image #1. Create video start frame at image #2. Model wearing pajamas as the reference image #2 with background remains consistent and undistracting. The model gently stretches her arms upwards, taking a soft, natural breath. Her expression is relaxed and sleepy. The movement is smooth and fluid, highlighting the comfortable fit and fabric of the pajamas. The video must maintain the exact same face, identity, and outfit as the reference image. Smooth, natural, and highly realistic motion. High quality, 8k resolution, cinematic lighting. Music background: No audio track is required for this video. |
3.2. Cảnh mở hộp
| Hình ảnh | Prompt ảnh |
 |
Messy hair, no makeup, wearing pajamas, holding the package with both hands. The package MUST look EXACTLY like the provided reference package image. She subtle smile, showing a calm and slightly curious expression. Keep the face and pajamas from the reference image exactly the same, but change the arm pose to hold the package. |
| Prompt video |
| Use the same face identity from reference image #1. The model is wearing the same pajamas as in reference image #3. The outfit she is wearing must remain unchanged throughout the video. The model gently holds a shipping box in reference image #3 in front of her. She slowly opens the box with a gentle, subtle smile, showing a calm and slightly curious expression. Inside the box is the garment from reference image #2. Keep the garment from the reference image exactly the same. She does not take the garment out. Instead, she gently touches and slightly lifts a part of the dress while it remains inside the box, allowing the fabric to move softly and reveal its texture. Her movements are slow, natural, and delicate. Her hands move smoothly, creating subtle, realistic folds in the fabric inside the box. Maintain exact facial identity, consistent proportions, and stable appearance. The worn outfit must not change or be replaced. Lighting & environment: Soft natural lighting with gentle shadows. Stable lighting with minimal variation. Background remains clean and undistracting. Style: Photorealistic, lifestyle video, cinematic realism, shallow depth of field. Audio: No audio track is required for this video. Technical: 4K UHD, 60FPS, smooth motion. Constraints: No outfit change, no morphing, no distortion, no flicker, no extra fingers, no anatomical errors. Safety: Neutral, non-sexualized presentation, no suggestive posing. |
3.3. Cảnh xem sản phẩm
| Hình ảnh | Prompt ảnh |
 |
Use the same face identity from reference image #1. Scene set inside an apartment bedroom, matching the environment and style of reference image #3. A young woman is sitting on the edge of the bed. She is wearing pajamas, with messy, tousled hair and no glasses. She is holding the outfit from reference image #2 in her hands. The atmosphere is natural and intimate, with a realistic bedroom setting. Soft lighting, slightly casual and unposed feeling. The bed, sheets, and surrounding furniture should match the style and layout of image #3. Her posture is relaxed, slightly leaning forward as she looks at the outfit in her hands. |
| Prompt video |
| Use the face identity from reference image #1. The garment in reference image #2 is the same item held by the model. The model is wearing the same pajamas as in reference image #3. The video must start exactly from reference image #3, where the model is already wearing pajamas and holding the garment. The starting frame must closely match reference image #3 in composition, pose, clothing, and background. The model gently lifts the garment slightly and examines it with a soft, thoughtful expression. She slowly turns the clothing from side to side, making small, natural adjustments to observe the details. Her movement is subtle and controlled, with slight breathing and minimal body motion. Camera remains stable with a gentle cinematic feel, using a medium shot and a very subtle push-in. Depth of field is shallow, keeping focus on the model and the garment. No audio track is required for this video. Lighting is soft and natural with gentle shadows. Light remains stable with minimal variation. Background stays consistent and undistracting. Style: photorealistic, lifestyle, cinematic realism, natural human motion. Quality: 4K UHD, 60fps, smooth motion, consistent identity, no flicker, no distortion, no morphing, accurate anatomy. Safety: neutral, non-sexualized presentation, natural everyday movement, no suggestive posing. |
3.4. Cảnh chuyển outfit mẫu mặc sản phẩm
| Hình ảnh | Prompt ảnh |
 |
Full-body shot of the young woman from reference image #1, with her signature long wavy hair and elegant makeup, standing naturally in the bright apartment from image #3. She is positioned in front of a clean, empty bed with her arms relaxed at her sides. She is wearing the exact outfit from image #2, maintaining the specific fabric texture and design. The background must be completely clear of any other people or white pajama fragments. The soft natural morning light from the window must interact realistically with her skin and the black fabric, casting accurate shadows on the floor and bed behind her. Photorealistic, 8k resolution, cinematic lighting, seamless environment integration, high fidelity, professional photography style. |
| Prompt video |
| Use the same face identity from reference image #1. Outfit as the reference image #2. Create video start frame at image #3 with background remains consistent and undistracting. Preserve the same person, facial identity, body proportions, outfit, hairstyle, and overall appearance. Maintain consistent identity across all frames with no changes or distortion. Create a realistic vertical video, duration 8 seconds, aspect ratio 9:16. Scene: A calm indoor setting consistent with the reference image. The subject is standing and begins a gentle, natural walk within the space. The environment remains unchanged and realistic. Motion (natural walking – safe & neutral): The woman performs a slow, natural walk in a relaxed, everyday manner. Movement is balanced and realistic, with natural weight shifting between steps. Arms swing gently in a subtle and natural way. Hands remain relaxed with minimal, non-expressive movement. Posture remains upright and composed, like a casual lifestyle or fashion walk. Facial expression stays neutral and calm, with occasional natural blinking. Camera: Camera remains mostly stable with a very slight forward or backward drift. Framing stays consistent at a respectful medium distance (no extreme close-up). No emphasis on specific body parts; full-body or mid-shot framing only. Smooth continuous motion with no sudden zoom or focus shifts. Lighting & environment: Soft natural lighting with gentle shadows. Light remains stable with only subtle natural variation. Background remains consistent and undistracting. Music background: No audio track is required for this video. Style: Photorealistic, lifestyle video, natural human movement, cinematic realism, shallow depth of field. Quality constraints: 4K UHD, 60FPS, smooth motion, consistent identity, no flicker, no distortion, no morphing, no extra fingers, no anatomical errors. Safety constraints: Non-sexualized, neutral presentation, everyday movement, no suggestive posing, no emphasis on body details.” |
3.5. Cảnh catwalk & tạo dáng
✔ High angle (từ trên xuống)
| Hình ảnh | Prompt ảnh |
 |
HIGH ANGLE SHOT (Bird’s-eye view). The camera is positioned high above the model’s head, looking down. Framing from the top of the head down to the calves. MANDATORY: Keep the exact same outfit and face. The perspective MUST clearly show a high-angle downward view. |
| Prompt video |
| Use the same face identity from reference image #1. Outfit as the reference image #2. Create video start frame at image #3 with background remains consistent and undistracting. The model slowly looks up towards the camera with a confident, gentle smile. She shifts her weight slightly, causing the outfit to drape naturally. The movement highlights the overall silhouette of the clothing from a high-angle perspective. The video must maintain the exact same face, identity. Smooth, natural, and highly realistic motion. High quality, 8k resolution, cinematic lighting. Lighting & environment: Soft natural lighting with gentle shadows. Light remains stable with only subtle natural variation. Style: Photorealistic, lifestyle video, natural human movement, cinematic realism, shallow depth of field. Music background: No audio track is required for this video. Quality constraints: 4K UHD, 60FPS, smooth motion, consistent identity, no flicker, no distortion, no morphing, no extra fingers, no anatomical errors. Safety constraints: Non-sexualized, neutral presentation, everyday movement, no suggestive posing, no emphasis on body details. |
✔ Low angle (từ dưới lên)
| Hình ảnh | Prompt ảnh |
 |
LOW ANGLE SHOT (Worm’s-eye view). The camera is positioned low near the ground/thigh level, looking up at the model. Full body framing from the feet up to the top of the head. MANDATORY: Keep the exact same outfit and face. The perspective MUST clearly show a low-angle upward view, making the model look tall and powerful. |
| Prompt video |
| Use the same face identity from reference image #1. Outfit matches reference image #2. The scene begins from reference image #3 with a clean, consistent, and undistracting indoor background. A young woman stands in a relaxed posture and slowly turns her shoulders slightly to the side. As she moves, the fabric of her outfit flows naturally, subtly revealing its structure and premium texture. Her expression remains calm and neutral, with gentle, controlled movement. Camera uses a low-angle perspective, slowly pushing in to enhance presence and elegance. Motion is smooth, natural, and continuous with no abrupt changes. No audio track is required for this video. Lighting is soft and natural with gentle shadows. Light remains stable with minimal variation. Background stays consistent and simple. Style: photorealistic, lifestyle, cinematic realism, shallow depth of field. Quality: 4K UHD, 60fps, smooth motion, consistent identity, no flicker, no distortion, no morphing, accurate anatomy. Safety: neutral, non-sexualized presentation, natural everyday movement, no suggestive posing. |
✔ Back shot
| Hình ảnh | Prompt ảnh |
 |
BACK SHOT (View from behind). The camera is positioned behind the model, showing the back of the outfit and hair. Low angle camera looking slightly up. Full body framing from feet to the top of the head. MANDATORY: Keep the exact same outfit. The perspective MUST clearly show the model from behind. |
| Prompt video |
| Use the same face identity from reference image #1. Outfit as the reference image #2. Create video start frame at image #3 with background remains consistent and undistracting.. The model walks slowly and gracefully away from the camera. The movement highlights the back details, flow, and drape of the outfit. Her hair sways gently with each step. The video must maintain the exact same face, identity, and outfit as the reference image. Smooth, natural, and highly realistic motion. High quality, 8k resolution, cinematic lighting. Lighting & environment: Soft natural lighting with gentle shadows. Light remains stable with only subtle natural variation. Background remains consistent and undistracting. Style: Photorealistic, lifestyle video, natural human movement, cinematic realism, shallow depth of field. Music background: No audio track is required for this video. Quality constraints: 4K UHD, 60FPS, smooth motion, consistent identity, no flicker, no distortion, no morphing, no extra fingers, no anatomical errors. Safety constraints: Non-sexualized, neutral presentation, everyday movement, no suggestive posing, no emphasis on body details. |
✔ Medium view
| Hình ảnh | Prompt ảnh |
 |
MEDIUM SHOT (Shoulder to knee). The camera frames the model from the shoulders down to the knees. Focus on highlighting the patterns on the outfit. The model is gently touching the hem of the outfit to guide the viewer’s eye. MANDATORY: Keep the exact same outfit and face. |
| Prompt video |
| Use the same face identity from reference image #1. Outfit matches reference image #2. The scene begins from reference image #3, maintaining a consistent and simple indoor bedroom background. A young woman stands naturally and gently touches the hem of her dress, softly sweeping the fabric with her fingers. The motion is slow, graceful, and natural, highlighting the lightness and subtle texture of the fabric. She slightly shifts her posture while maintaining a calm and neutral expression. Camera remains stable with a slight cinematic handheld feel. Medium shot transitioning subtly to a closer framing. Shallow depth of field keeps focus on the subject. No audio track is required for this video. Lighting is soft and natural, with gentle shadows. No sudden lighting changes. Background remains consistent and undistracting. Style: photorealistic, lifestyle, cinematic realism, natural human motion. Quality: 4K UHD, 60fps, smooth motion, consistent identity, no flicker, no distortion, no morphing, accurate anatomy, no extra fingers. Safety: neutral, non-sexualized presentation, natural everyday movement, no suggestive posing. |
✔ Close-up
| Hình ảnh | Prompt ảnh |
 |
CLOSE-UP SHOT (Detail shot of the outfit). The camera frames from the shoulder to the knee, highlighting the patterns on the outfit. Focus on the fabric details. MANDATORY: Keep the exact same outfit. |
| Prompt video |
| Use the face identity from reference image #1. The outfit matches reference image #2. The video must start exactly from reference image #3, where the model is already wearing the outfit. The starting frame should closely match reference image #3 in pose, composition, clothing, and background. A close-up shot gently captures the fabric as the model remains mostly still. She shows subtle natural breathing, creating slight, realistic movement in the clothing. Her posture is relaxed with a calm, neutral expression. The camera performs a very slow and minimal lateral movement, briefly observing the fabric texture before settling naturally on the upper body. Motion remains subtle and controlled. Lighting is soft and natural with gentle shadows. Light remains stable with minimal variation. Background stays clean and undistracting. Style: photorealistic, natural, shallow depth of field. Quality: 4K UHD, 60fps, smooth motion, consistent identity, no flicker, no distortion, no morphing, accurate anatomy. No audio track is required for this video. Safety: neutral, non-sexualized presentation, natural everyday movement, no suggestive posing. |
4. Tips tối ưu khi dùng bộ prompt
Để đảm bảo chất lượng video ổn định và chuyên nghiệp, bạn nên luôn sử dụng hệ thống reference image rõ ràng, trong đó image #1 đại diện cho khuôn mặt nhân vật, image #2 là outfit và image #3 là bối cảnh. Việc phân tách vai trò từng ảnh giúp AI hiểu đúng và giữ được sự nhất quán xuyên suốt các cảnh.
Bên cạnh đó, cần hạn chế thay đổi quá nhiều giữa các scene, đặc biệt là về ánh sáng, background và tỉ lệ cơ thể, vì đây là những yếu tố quyết định độ liền mạch của video. Khi thiết kế chuyển động, hãy ưu tiên các hành động nhẹ nhàng, tự nhiên như cử chỉ tinh tế hoặc chuyển động chậm, vì AI xử lý những dạng này tốt hơn rất nhiều so với các chuyển động nhanh hoặc phức tạp.
Cuối cùng, luôn đảm bảo prompt có các constraint quan trọng như “no flicker, no distortion, no morphing” để tránh lỗi hình ảnh, méo hình hoặc biến dạng nhân vật trong quá trình render.
5. Ứng dụng thực tế
Bạn có thể dùng bộ prompt này để:
- Bán quần áo online
- Làm video TikTok affiliate
- Tạo content brand cá nhân
- Xây dựng thư viện workflow bán trên website
