VEO3 TikTok
视频脚本生成手册

第1步：用户发产品名称 + 产品图
第2步：出一份确认表，逐项确认
第3步：确认后出完整脚本

类型一：产品是主角

0-2s→场景一/台词4-6词 · 2-4s→场景二/开箱/4-6词
4-6s→核心功能/4-6词 · 6-8s→体验/6-8词
8-10s→多场景快切/5-7词 · 10-12s→蒙太奇/3-5词
12-15s→人物+产品出镜收尾/10-12词 · 总计约40词

类型二：人是主角

0-2s→痛点展示/6-9词 · 2-4s→产品开箱/4-6词
4-6s→开始使用/4-6词 · 6-8s→过程特写/4-6词
8-10s→效果展示/4-6词 · 10-12s→对镜展示/5-8词
12-15s→收尾促单/5-8词 · 共计约40词 · 人物全程出镜

类型三：共创

0-2s→准备/倒入/5-7词 · 2-4s→产品工作/3-5词
4-6s→成品特写/4-7词 · 6-8s→快切/4-6词
8-10s→价值感/5-7词 · 10-12s→体验ASMR/0词
12-15s→收尾/7-10词 · 共计约35-40词

在3大类基础上精确定位：

人物A（主力）

A young woman with long wavy dark brown hair past
her shoulders. Warm skin tone with a natural glow.
Dark brown eyes. Slim athletic build. Wearing an
oversized faded gray crewneck sweatshirt. Hair
loosely tucked behind one ear.

人物B（黑人女性）

A real American woman with natural dark coily hair
loosely gathered or free around her face, deep rich
brown skin with a warm healthy glow, dark brown
eyes with a relaxed and genuine expression, a
naturally athletic build, a delicate thin gold chain
necklace and small gold hoop earrings.

人物C（拉丁裔女性）

A real American woman with long dark straight hair,
olive warm skin tone with a natural healthy glow,
dark expressive eyes, a naturally proportioned
build, a delicate thin gold chain necklace.

• 只改变：服装+场景+背景
• 外貌描写100%固定不变

第一轮：产品调性

第二轮：场景匹配

• 背景可以不完美，路人/停着的车没关系
• 手持轻微晃动·自然光·不完美构图
• 要给人物一个"触发她拍视频的理由"

0-4s   第一镜 — HOOK（稍长，给观众进入状态）
4-11s  第二镜 — 核心展示（最长，产品高潮时刻）
11-15s 第三镜 — 结尾 CTA（简短干净）

1. 节奏快慢？ 慢节奏→插入镜跟着慢甚至不需要 / 快节奏→可强化转折

2. 情绪转折需要反应镜"告诉"观众吗？ 画面已震惊→不需要 / 画面平静→可帮助传递

3. 衔接自然吗？ 自然过渡→可加 / 突然闪出→坚决不加

产品移动时

• 写出人的手握在哪个部位
• 手和手臂必须在画面中
• 远镜头→全身可见 / 近镜头→至少保留人手

产品静止但功能运行

• 允许纯产品画面
• 但前一镜头人刚放下，后一镜头人要回来
• 不要让产品"自己工作"超过3秒

三步检查

1. 这个画面里有没有人？
2. 产品在移动→人的手握在哪里？
3. 检查连续性：人不能在镜头A出现、B消失、C又出现

前景层：女人声音，连续不断
背景层：产品音效+环境音，同时播放
两层共存从头到尾

台词生成步骤

1. 叙事弧：初始状态→发生什么→结果
2. 核心句：最强那句话（通常即Hook）
3. 填充前后：保持35-42词
4. 共鸣点：有没有"每个人都会共鸣"的句子？
5. 朗读删除：所有不像口语的词

金句模板

• ✅ "Wait. WAIT."（她已在水里直接反应）
• ✅ "Look at my ceiling right now."
• ✅ "Winter destroyed my yard. Look at this."
• ✅ "She came over to swim. That was 3 hours ago."
• ❌ "Okay so I just got this and wanted to show you..."

0-3s:   正常说话 — 建立人物，痛点钩子（产品音效背景）
3-11s:  正常说话 — 核心展示，台词可减少但不消失（音效持续）
11-15s: 正常说话 — 干净收尾，一句话CTA（音效持续）

LANGUAGE LOCK: This video must be generated entirely
in English. The character speaks only natural American
English throughout. Zero Chinese dialogue, zero Chinese
subtitles, zero Chinese text of any kind anywhere
in the video.

No subtitles, no on-screen text, no stickers,
no background music, no color grading filters
applied in post, and a natural ending like
a real TikTok post.

--ar 9:16

通用防变形框架

PRECISE PRODUCT STRUCTURE — MUST MATCH EXACTLY:
[详细描述每个零件、连接方式、材质]

CRITICAL DETAIL — [最容易变形的部分]:
[极度详细描述该部分物理特征]

Do NOT generate [错误形态1].
Do NOT generate [错误形态2].
The structure is exactly [正确描述].

镜头风险

① 面部反应 — 闭眼满足感/睁眼震惊/嘴角smug
② 手触产品 — 手指按压弹回/手掌贴合/水珠滚动
③ 水面/光线 — 手指拨水涟漪/阳光折射/水珠滚落
④ 印花面料 — 极近图案细节/光线纹理
⑤ 道具ASMR — 饮料吸管/冰块碰撞/瓶盖打开

评论区引爆点

• "this is literally me"→共鸣 · "@朋友"→传播 · "where did you get this"→好奇
• 说一件"每个人都做过但从没人说出来"的事
• 制造信息缺口/"Watch."/结局不完整让观众猜

• 每段间加空行，台词单独一行
• 产品/场景/画面描述分段落
• 不整段挤在一起
• 平台限字数：去空行压缩但保留结构

An ultra-realistic TikTok-style vertical video in
9:16 ratio, exactly 15 seconds. All dialogue is in
natural American English. No Chinese text anywhere.

The video features adhesive bra inserts that match
the provided reference image exactly. Same shape,
same color, same material, same proportions. The
inserts are nude skin-tone silicone pads with a
curved triangle fan shape resembling a natural cup
contour. The edges are thin and the center is
slightly thicker for volume. The surface is smooth
and skin-like with a matte finish that mimics real
skin texture. The back surface has a gentle adhesive
layer covered by a thin protective film that peels
off before use.

The scene is a real small apartment bedroom with a
full-length mirror propped against the wall. Soft
natural daylight coming through thin white curtains
on a window to the side. The bed has a wrinkled
white duvet pushed to one side with a few summer
tops laid out flat on the other side. A small
wooden dresser sits beside the mirror with everyday
items scattered on top. A hairbrush. A half-empty
water bottle. A small jewelry dish with a couple of
rings. The room looks lived in and real. This is not
a commercial. This is someone getting dressed on a
normal morning.

A woman is present throughout the entire video. She
is the focus. The product is used on her. Her voice
is the foreground layer running continuously from
start to finish. Room ambient sounds and product
sounds play simultaneously underneath her voice at
all times as the background layer.

0 to 2 seconds. The woman stands in front of the
full-length mirror wearing a white summer top. The
fabric does not sit right. It gaps at the neckline
and pulls awkwardly across the chest area. The
shoulder straps sit uneven. The overall look is
messy and unflattering. She looks at herself in the
mirror. She tugs at the neckline trying to adjust
it. It does not help. She lets go and sighs. She
says I love this top. It never fits right.

2 to 4 seconds. Close-up of her hands picking up
the nude skin-tone curved silicone inserts from the
bed where they sit next to the other summer tops.
She holds one insert up near the window so the
natural light catches the smooth matte surface. The
curved triangle shape is clearly visible. The nude
color matches her skin tone closely. She turns it
over and peels the thin protective film off the
sticky back surface. The adhesive layer underneath
is slightly glossy. The protective film comes off
with a soft satisfying peel sound. She says No
sewing. No tape. Just these.

4 to 6 seconds. Quick cut transition. The woman now
has the inserts on. She pulls the same white summer
top on over her head. She smooths the fabric down
with both hands from the shoulders to the waist.
The fabric now sits completely differently. Clean
flat neckline. Even shoulder straps. Smooth fabric
across the chest with no gaps and no pulling. The
difference from the first shot is dramatic. She
looks at herself in the mirror and tilts her head
slightly. A small satisfied expression. She says
They just work.

6 to 8 seconds. She stands in front of the full-
length mirror. Same white top. She turns slowly to
the side. The fabric drapes smoothly with no
bunching and no visible lines. She turns to the
other side. Same clean result. She faces forward
again. The silhouette looks polished and put
together. The white fabric catches the natural
window light evenly across the surface. No shadows
from edges. No bumps. Just smooth clean fabric. She
says Look at the difference.

8 to 10 seconds. Quick montage cuts. Same woman
different outfits. Shot one. A flowy sundress with
thin straps. The dress hangs perfectly with smooth
clean lines across the chest. Shot two. A silk
camisole in a soft pink tone. The fabric drapes
naturally without any visible lines underneath. Shot
three. A basic white tank top. Clean fitted
silhouette with no bumps and no gaps. Each outfit
looks intentionally styled and put together. The
woman looks confident in each one. She says Every
top in my closet.

10 to 12 seconds. Back to the white summer top. She
stands in front of the mirror. She raises both arms
above her head in a full stretch. The top stays
exactly in place. No shifting. No riding up. No
adjusting needed. She puts her arms down. The top
is still perfectly positioned. She leans forward
slightly then straightens back up. Everything stays
exactly where it should. She says It does not budge.

12 to 15 seconds. Front-facing selfie. A young
woman with long wavy dark brown hair falling past
her shoulders with a slight natural volume at the
crown and loose waves through the ends. Warm medium
skin tone with a natural healthy glow. Expressive
deep brown eyes with naturally full dark lashes.
Slim and naturally proportioned athletic build.
Wearing the white summer top which now fits
perfectly with clean lines and a polished
silhouette. The full-length mirror is visible
behind her showing her full outfit. She speaks to
camera casual half-laughing like she is sending a
voice message to a friend. She says Summer is
coming and I am ready. She reaches down and picks
up a small crossbody bag from the bed. She slings
it over her shoulder and walks toward the bedroom
door. The camera stays on the mirror as she walks
out of frame. Her reflection lingers for a moment
in the empty mirror. Then the video hard-cuts to
black.

Sound design throughout. The woman's voice is the
foreground layer running continuously from start to
finish. The background layer is fabric rustling as
she pulls tops on and off, protective film peeling,
fabric smoothing over body, mirror ambient room
sound, footsteps as she walks toward the door. Both
layers play simultaneously at all times. No
subtitles. No on-screen text. No stickers. No
background music. No color grading filters. Natural
ending like a real TikTok post.

An ultra-realistic TikTok-style vertical video in
9:16 ratio, exactly 15 seconds. All dialogue is in
natural American English. No Chinese text anywhere.

The video features a push lawn sweeper that matches
the provided reference image exactly. Same frame
shape, same color, same brush configuration, same
hopper bag design, same wheel design, same
proportions. The sweeper has a black frame with
orange accent knobs, orange wheel hubs, and orange
brand markings. It has a U-shaped curved handle for
comfortable pushing grip. Two spinning brushes are
mounted underneath the frame. A mesh collection
hopper bag is mounted on the back. Large rubber
wheels with tread grip are on each side.

The scene is a real suburban house front yard in
early spring. The lawn is patchy with some brown
dead spots from winter. Fallen small branches, dry
leaves leftover from winter, scattered twigs, and
small debris are scattered across the grass. Some
early green shoots are starting to appear. The
flower beds along the house have old dried leaves
piled in them. The sky is overcast with soft gray
spring light. A rake leaning against the garage wall
unused. This is not a commercial. This is someone's
actual yard on a weekend morning in early spring.

A woman narrates the entire video in a casual
conversational tone like she is talking to a
neighbor while working in her yard. Her voice is the
foreground layer running continuously from start to
finish. Product sounds and outdoor ambient sounds
play simultaneously underneath her voice at all
times as the background layer.

0 to 2 seconds. Wide shot of the front yard covered
in winter mess. Dead leaves, small sticks, random
debris across the lawn. The grass underneath is
patchy brown and green. The flower beds have old
dried leaves piled up. The rake leans against the
garage wall untouched. She says Winter destroyed my
yard. Look at this mess.

2 to 4 seconds. Low-angle shot at ground level. The
lawn sweeper sits at the edge of the yard ready to
go. The black frame with orange accents is visible.
The two spinning brushes are visible underneath. The
mesh hopper bag hangs empty on the back. The
U-shaped curved handle rises up from the frame. A
woman's hands enter the frame from above and grip
the U-shaped handle. Her faded flannel sleeve is
visible at the edge of frame. She says Then I
remembered I bought this.

4 to 6 seconds. Medium shot from the side. The
woman is visible from the waist down pushing the
sweeper forward. Her hands grip the U-shaped curved
handle firmly. Her legs walk in steady strides. The
two spinning brushes rotate rapidly underneath the
frame. Dry leaves, small twigs, and debris swirl and
funnel upward into the mesh hopper bag. Camera
follows alongside showing the woman pushing and the
brushes spinning simultaneously. The path behind the
sweeper is immediately clean grass. The contrast
between the messy lawn ahead and the clean lawn
behind is dramatic. Her gloved hands adjust grip on
the handle as she turns slightly. ASMR of spinning
brushes and rustling debris. She says You just push
it. That is it.

6 to 8 seconds. Close-up of the spinning brushes
from underneath. Leaves and small twigs get caught
in the bristles and funneled into the collection
chute leading to the hopper bag. Satisfying
mechanical spinning sound. Camera tilts up to show
the mesh hopper bag filling up visibly with debris.
The orange wheel hub catches light as the wheel
rolls. At the edge of frame the woman's hand is
still gripping the handle and the sweeper is still
moving forward. The bag is getting fuller with each
pass. She says It picks up everything. Leaves.
Sticks. Everything.

8 to 10 seconds. Wide shot from the driveway. The
woman pushes the sweeper in straight rows across the
lawn. Her full body is visible from behind. Faded
flannel shirt. Messy ponytail. She leans into the
handle with each push. Each pass leaves a perfect
stripe of clean grass. The black frame and orange
accents of the sweeper contrast against the green
and brown lawn. The yard is transforming from fully
covered in winter mess to half clean. The hopper bag
on the back is now visibly stuffed full. She says
Look at my yard right now. Half of it is done.

10 to 12 seconds. Low-angle tracking shot. The
sweeper rolls past camera at close range. The large
rubber wheels with tread grip the uneven ground
smoothly. The orange wheel hubs spin. The spinning
brushes whir. Debris disappears into the bag. The
strip of clean grass behind is starting to show
early spring green. Camera stays low and follows the
sweeper as it moves across the yard. At the top of
frame the woman's hands are visible pushing the
handle and her feet are walking alongside. She says
The wheels go over everything. Bumps. Sticks. Wet
leaves. No problem.

12 to 15 seconds. Front-facing selfie. A young
woman with long wavy dark brown hair pulled back in
a messy ponytail. Warm skin tone with a natural
glow. Dark brown eyes. Slim athletic build. Wearing
a faded flannel shirt with the sleeves rolled up and
old jeans. A small twig stuck in her ponytail. She
looks slightly out of breath but satisfied. She
holds the sweeper handle in one hand. The clean yard
is visible behind her. The black frame with orange
accents catches light. She speaks to camera casual
half-laughing. She says I was about to pay someone
two hundred bucks. I just did it myself. My back
feels fine by the way. That is new. She turns the
camera around to show the clean yard behind her.
Then back to herself. She shrugs. The video hard-
cuts to black.

Sound design throughout. The woman's voice is the
foreground layer running continuously from start to
finish. The background layer is spinning brush
mechanical whir, rustling of dry leaves and twigs
being swept, rubber wheels rolling over grass and
gravel, debris hitting the mesh bag, distant birds,
spring breeze. Both layers play simultaneously at
all times. No subtitles. No on-screen text. No
stickers. No background music. No color grading
filters. Natural ending like a real TikTok post.

LANGUAGE LOCK: This video must be generated entirely
in English. The character speaks only natural American
English throughout. Zero Chinese dialogue, zero Chinese
subtitles, zero Chinese text of any kind anywhere
in the video.

Create an ultra-realistic TikTok-style vertical video (9:16),
exactly 15 seconds long, based strictly on the provided
product image of [产品名称] — [产品物理描述，极度详细].

[如有变形风险，加入PRECISE PRODUCT STRUCTURE段落]

The video features [人物外貌描写].
[服装描述 — 碰巧穿着感].
[人物状态 — 她在做什么，为什么拍这个视频].

Filmed [场景 — 真实UGC感].
[氛围定性句 — "This is not a commercial. This is..."].

[音频 — 前景人声+背景产品音效，两层同时播放]

0–4s HOOK: [第一镜 — 最强画面，第一句即结论]

4–11s CORE DEMO: [核心展示 — 动作连续，台词简短]

11–15s ENDING: [收尾 — 风格匹配CTA]

No subtitles, no on-screen text, no stickers,
no background music, no color grading filters
applied in post, natural ending like a real TikTok post.

--ar 9:16