street-interview-ad

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Street Interview Ad

街头采访广告

Purpose

目的

Create or audit ads that use a believable public conversation to move from curiosity or skepticism to grounded explanation, proof, and an optional CTA.

Treat the format as a production grammar, not permission to fabricate customers, testimonials, experts, or spontaneous public reactions.

创建或审核以可信的公开对话为载体的广告，引导受众从好奇或怀疑转向清晰的解释、真实证据，以及可选的CTA（行动号召）。

将该格式视为制作准则，而非虚构客户、推荐语、专家或自发公众反应的许可。

Choose The Mode

选择模式

Create Mode

创建模式

Use when the user supplies an offer, brand brief, product facts, approved claims, rough concept, or script and wants a street-interview direction.

Produce a

Street Interview Grammar Plan

, shot script, production locks and checks, one approval checkpoint, and direct handoff after approval.

当用户提供优惠方案、品牌简报、产品事实、已获批的宣传点、初步概念或脚本，并希望采用街头采访方向时使用此模式。

产出一份

街头采访准则方案

、拍摄脚本、制作锁定项与检查项、一个审批节点，获批后直接交付制作。

Review Mode

审核模式

Use when the user supplies an existing ad and asks whether it fits this grammar or how it should improve.

Use

$video-analysis

first to obtain observable shot and audio evidence. Then return:

factual structure, hook, product timing, proof, and ending status
a timecoded
```
Strong / Partial / Weak / Missing
```
audit
truth-status risks, meaningful deviations, and precise amendments

Do not rewrite the ad or enter production unless the user asks.

当用户提供现有广告，并询问其是否符合该准则或如何改进时使用此模式。

首先使用

$video-analysis

获取可观察的镜头和音频证据。然后返回：

事实结构、钩子、产品出场时机、证据以及结尾状态
带时间码的
```
达标/部分达标/未达标/缺失
```
审核结果
真实性风险、重大偏差以及具体修改建议

除非用户要求，否则不要重写广告或进入制作环节。

Create Workflow

创建工作流程

1. Ground The Brief

1. 锚定简报

Read the smallest useful source context:

offer facts, audience, brand voice, approved claims, and available proof
product, persona, location, UI, audio, or reference assets
prior ad analysis when explicitly relevant
duration, aspect ratio, model, resolution, and output preferences when supplied

Lock:

offer_type

physical_product | software | service | community | content

```
offer_capability
```
```
audience
```
```
specific_pressure
```
```
interview_premise
```
```
interviewer_role
```
```
respondent_role
```

respondent_truth_status

approved_customer | approved_expert | spokesperson | actor | illustrative_persona

```
proof_available
```
```
claim_boundaries
```
```
setting_and_format
```

Do not write first-person results, expert authority, public consensus, or spontaneous discovery unless the supplied evidence supports that truth status. Actors and illustrative personas must not be presented as real customers or independent passersby.

阅读最有用的最小范围源上下文：

优惠方案事实、受众、品牌调性、已获批的宣传点、可用证据
产品、人物角色、地点、UI、音频或参考素材
明确相关的过往广告分析
若提供，需包含时长、宽高比、模型、分辨率和输出偏好

锁定以下内容：

offer_type

physical_product | software | service | community | content

```
offer_capability
```
```
audience
```
```
specific_pressure
```
```
interview_premise
```
```
interviewer_role
```
```
respondent_role
```

respondent_truth_status

approved_customer | approved_expert | spokesperson | actor | illustrative_persona

```
proof_available
```
```
claim_boundaries
```
```
setting_and_format
```

除非提供的证据支持相应的真实性，否则不要撰写第一人称成果、专家权威、公众共识或自发发现的内容。演员和示例人物角色不得被呈现为真实客户或独立路人。

2. Design The Street Grammar

2. 设计街头采访准则

Read

references/street-grammar.md

and

references/continuity-and-proof.md

Produce a

Street Interview Grammar Plan

with exactly these fields:

```
core_human_state
```
```
specific_pressure
```
```
interview_premise
```
```
interviewer_role
```
```
respondent_role_and_truth_status
```
```
question_ladder
```
```
camera_language
```
```
framing_and_eyeline
```
```
movement_and_edit_grammar
```
```
microphone_behavior
```
```
product_entry
```
```
product_gesture
```
```
proof_plan
```
```
sound_world
```
```
text_and_subtitle_behavior
```
```
continuity_locks
```
```
ending_and_cta
```

State

Claim Boundaries

immediately after the plan. Do not force a fixed location, respondent count, product category, music track, hook box, or ending formula.

阅读

references/street-grammar.md

和

references/continuity-and-proof.md

。

产出一份包含以下确切字段的

街头采访准则方案

：

```
core_human_state
```
```
specific_pressure
```
```
interview_premise
```
```
interviewer_role
```
```
respondent_role_and_truth_status
```
```
question_ladder
```
```
camera_language
```
```
framing_and_eyeline
```
```
movement_and_edit_grammar
```
```
microphone_behavior
```
```
product_entry
```
```
product_gesture
```
```
proof_plan
```
```
sound_world
```
```
text_and_subtitle_behavior
```
```
continuity_locks
```
```
ending_and_cta
```

在方案后立即说明

宣传边界

。不要强制固定地点、受访者数量、产品类别、音乐曲目、钩子框架或结尾公式。

3. Write The Shot Script

3. 撰写拍摄脚本

Read

references/shot-script-rubric.md

Return a shot table with exactly these columns:

time | visual | framing/composition | camera motion | subject/action | product/prop relationship | lighting/color | edit | sound/music | speech/text | purpose

After the table, provide:

```
Continuity Locks
```
```
Proof Status
```
```
Claim Boundaries
```
```
Generation Risk Check
```

Keep one dominant action and one conversational job per row. Write all spoken lines needed for timing, but keep final subtitles, hook overlays, package text, UI labels, legal copy, and CTA graphics for deterministic post-production.

阅读

references/shot-script-rubric.md

。

返回包含以下确切列的镜头表格：

时间 | 视觉内容 | 取景/构图 | 镜头运动 | 主体/动作 | 产品/道具关联 | 灯光/色彩 | 剪辑 | 声音/音乐 | 台词/文字 | 目的

表格之后提供：

```
连续性锁定项
```
```
证据状态
```
```
宣传边界
```
```
生成风险检查
```

每行保留一个核心动作和一项对话任务。撰写所有用于计时的台词，但将最终字幕、钩子叠加层、包装文字、UI标签、法律文案和CTA图形留待确定性后期制作。

4. Ask For Approval

4. 请求审批

Ask one concise approval question in the user's language. English default:

Does this street-interview direction feel right? If yes, I will hand it directly to production.

Stop before production until the user approves.

用用户的语言提出一个简洁的审批问题。默认英文表述为：

Does this street-interview direction feel right? If yes, I will hand it directly to production.

（中文：这个街头采访方向是否合适？若合适，我将直接交付制作。）

在用户批准前停止制作环节。

5. Handoff On Approval

5. 获批后交付

Read

references/workflow-handoff.md

and immediately route the locked plan and script.

Default to

$image-batch-runner

first when two-person continuity, recurring respondents, microphone/hand geometry, product handling, first frames, or street plates need stable assets. Feed approved stills into

$video-batch-runner

Go directly to

$video-batch-runner

when persona, product, microphone, scene, voice, and proof references are already stable and validated.

Do not output a separate production brief and do not ask a second generation confirmation.

阅读

references/workflow-handoff.md

，立即将锁定的方案和脚本交付。

当需要稳定素材来呈现双人连续性、重复受访者、麦克风/手部几何形态、产品操作、首帧或街景镜头时，默认优先使用

$image-batch-runner

。将获批的静态素材导入

$video-batch-runner

。

当人物角色、产品、麦克风、场景、声音和证据参考已稳定并验证时，直接使用

$video-batch-runner

。

无需输出单独的制作简报，也无需再次确认生成指令。

Quality Bar

质量标准

Open with a concrete question and show a visible listening state.
Keep microphone, eyeline, hand ownership, and interviewer geometry credible.
Build a question ladder before one legible product action and proof response.
Preserve handheld public-place dialogue with restrained edits and music.
End on a changed human state; add near-verbatim captions in post.

以具体问题开场，并展示可见的倾听状态。
确保麦克风、视线、手部归属和采访者位置关系可信。
在呈现清晰的产品动作和证据回应前，构建问题阶梯。
保留手持拍摄的公共场所对话风格，剪辑和音乐克制使用。
以受众状态的改变收尾；后期添加近乎逐字的字幕。

Non-Imitation And Truth Boundary

非模仿与真实性边界

Do not copy source faces, voices, exact locations, wardrobe combinations, brands, domains, signature lines, or distinctive background props.
Do not claim that an actor is a random passerby, customer, expert, or independent reviewer.
Do not invent medical, financial, comparative, rapid-result, performance, scarcity, discount, review, or guarantee claims.
Keep critical text out of generated frames and translate references into generic camera, movement, sound, state, proof, and edit attributes.

不得复制源素材中的面部、声音、精确地点、服装搭配、品牌、域名、标志性台词或独特背景道具。
不得声称演员是随机路人、客户、专家或独立评论者。
不得编造医疗、金融、对比、快速见效、性能、稀缺性、折扣、评价或担保类宣传点。
避免在生成的画面中出现关键文本，将参考素材转化为通用的镜头、运动、声音、状态、证据和剪辑属性。