AD: 5S CLIP PLAN: REFERENCE PLAN: <Picture 1> is the supplied Avatar visual identity reference for <Subject 1> (S1), a young adult female creator; she is visible on camera. <Picture 2> is the authoritative raw product reference for <Subject 2>, the VACY Advanced Botanical Crème jar; preserve its clear jar, reflective gold lid, gold VACY logo, and exterior form exactly. <Subject 1> (S1), the same presenter defined by <Picture 1>, is the sole audible speaker in the clip. No anonymous narrator, voice actor, or handling hand speaks. Product continuity: Preserve the VACY clear cosmetic jar, reflective gold lid, and gold VACY logo exactly as shown in the raw product reference. [Shot 1] | 00:00.000–00:05.000 | medium close-up from a phone on a tripod at eye level, one arm-length from <Subject 1> seated at a bright home vanity; <Subject 2> rests upright and closed on the vanity beside her, logo facing camera; she looks from the jar to camera with a curious, approving smile, then lightly points toward the jar without touching it, ending with the product clearly framed beside her | dialogue: <d>[English] I wanted a fragrance-free moisturizer, and this VACY crème has a really polished, simple formula.</d> | <Subject 1> (S1), on-camera lip sync, young adult female voice, subtle West Coast US accent, warm credible beauty-creator tone, moderate energy; natural steady conversational pace; same speaking rate across all clips; no rushed delivery, no elongated syllables, no dramatic slowdown. | locked tripod framing with a very slow two-inch digital push-in | soft window light, clean neutral vanity, temporally stable natural skin texture and low-contrast facial detail; no beauty filter or oily highlights | quiet room tone and a faint fabric shift as she settles. Frame: 9:16. Execute AD as one 5-second clip. - Render AD's ordered action, cast, hand/tool choreography, camera intent, cause/effect, dialogue, and end state. Do not add unrelated people, hands, props, scenes, claims, or beats. - Use each attached picture only for its AD REFERENCE PLAN role. Raw product pictures are authoritative for visible packaging, Logo, and exterior form; do not fabricate or morph them. - Do not show any model, person, face, body, or hands unless AD explicitly requires them. If AD has a HAND PLAN, render only its planned-action hands. Without one, do not add product-handling or focal hand action; passive hands of an AD-required visible person may remain natural, non-interacting, and away from the product. - Unless explicitly required by the AD or an attached reference image, no duplicate faces, bodies, twins, or body doubles. - Execute the speaker-subject bindings declared in AD exactly. For an off-screen speaker, retain the named `<Subject N>` and `(Sx)` from AD as the sole audible source; do not create an anonymous narrator, voice actor, or new speaker. Anonymous product-handling hands are silent unless AD explicitly assigns them a speaker ID. Preserve each speaker's exact AD-stated voice description. - DIALOGUE LOCK: For a non-silent clip, one or more `<d>[Language] ...</d>` blocks in source order are the complete final transcript and the only source of spoken words. `[Language]` is control metadata and must not be spoken. Speak every remaining word exactly once, in exactly that order. Do not paraphrase, translate, add, omit, repeat, abbreviate, or substitute any word, and do not use `<cutoff>` for a complete clip. Follow on-camera lip sync versus off-screen speech exactly - Clean footage: no added captions, subtitles, dialogue text, titles, labels, callouts, or overlays. Dialogue is audio-only. Do not visualize added on-screen text. Preserve only text printed on referenced product/packaging. No BGM. - Preserve AD timing, shot order, lighting, identity, product geometry, markings, and quantity.

AD: 5S CLIP PLAN: REFERENCE PLAN: <Picture 1> is the visual identity reference for <Subject 1> (S1), an adult Chinese female creator; <Subject 1> is visible on screen. <Subject 1> (S1) is the sole audible speaker. <Picture 2> is authoritative raw product evidence for the white JOMOO close-coupled toilet silhouette and finish. <Subject 2> is the traditional-style toilet shown only for the cleaning-frustration comparison. [Shot 1] | medium close-up from a phone camera at bowl height in a lived-in bathroom | <Subject 1> wears rubber gloves beside <Subject 2>, which starts with a visible rim groove; her right hand holds a toilet brush above the groove | her right hand scrubs the groove twice, the bristles catch at the seam, then she grimaces and releases the brush into an empty floor-side tray; the brush lies still and she looks to camera in frustration | dialogue: <d>[Chinese] 家人们，传统马桶这圈棱缝，真的能把人刷到崩溃。</d> | <Subject 1> (S1), on-camera lip sync, young adult female voice, natural Mainland Chinese, candid frustrated tone, brisk conversational pace | subtle handheld forward drift, slow and small | soft daylight bathroom, natural temporally stable skin texture and facial detail, no beauty-filter look | synchronized bristle scraping, a light plastic clatter, room tone [Shot 2] At 00:03.700 | tight punch-in on <Subject 1> and the discarded brush | the brush remains still in the tray and <Subject 1> points one gloved finger toward the troublesome groove | she holds the point for a beat as bold on-screen Chinese text bursts in: “刷马桶刷到崩溃？该换了！”; final state is her finger held near, not touching, the rim | dialogue: <d>[Chinese] 该换个好打理的了。</d> | <Subject 1> (S1), on-camera lip sync, young adult female voice, natural Mainland Chinese, candid frustrated tone, brisk conversational pace | quick digital punch-in, then locked | same bathroom and face-fidelity treatment | short text-impact hit and soft room tone Execute AD as one 5-second clip. - Use every attached picture only for the role assigned in AD's REFERENCE PLAN. Raw product references are authoritative for product identity, packaging, logo/text, color, shape, structure, quantity, orientation, and mechanism. - Do not show any model, person, face, body, or hands unless AD explicitly requires them. - Execute the speaker-subject bindings declared in AD exactly. For an off-screen speaker, retain the named `<Subject N>` and `(Sx)` from AD as the sole audible source; do not create an anonymous narrator, voice actor, or new speaker. Anonymous product-handling hands are silent unless AD explicitly assigns them a speaker ID. Preserve each speaker's exact AD-stated voice description. - Inside `<d>...</d>`, `[Language]` is control metadata and must not be spoken. Speak the remaining words exactly once without paraphrasing, translating, adding, omitting, or repeating. Follow on-camera lip sync versus off-screen speech exactly. - Generate clean footage for later subtitle compositing: no added captions, subtitles, dialogue text, titles, labels, callouts, or text overlays in any frame. Dialogue is audio-only. If AD mentions added on-screen text, do not visualize it. Preserve only text already printed on the referenced product or packaging. No BGM. - Preserve AD timing, shot order, cast, actions, camera, lighting, identity, product geometry, markings, quantity, and end state. Add no extra people, hands, props, scenes, or claims.

AD: 5S CLIP PLAN: REFERENCE PLAN: <Picture 1> is the supplied Avatar, the visual identity reference for recurring <Subject 1> (S1), visible on camera. <Subject 1> (S1) is a young adult male creator. <Picture 2> is the authoritative raw product-cover reference for <Subject 2>, the paperback book; preserve its exact visible cover artwork and text without fabrication or morphing. Product continuity: Keep <Subject 2> as the exact physical book shown in Picture 2, with its cover facing camera and supported upright on the desk. [Shot 1] | medium close-up from a tripod-mounted phone at eye level, facing <Subject 1> seated at a home-office desk; <Subject 2> stands cover-forward beside his elbow, starting fully visible | <Subject 1> leans slightly toward camera, glances from the book cover to the lens, then gives one small open-palm gesture beside the book without touching it; the book remains upright and cover-forward as the stable end state | dialogue: <d>[English] A Wall Street mystery with college recruiting, fundraising, and murder? That caught my attention.</d> | <Subject 1> (S1), on-camera lip sync, young adult male voice, neutral General American English with a subtle East Coast directness, candid and curious tone, moderate energy; natural steady conversational pace; same speaking rate across all clips; no rushed delivery, no elongated syllables, no dramatic slowdown. | locked tripod framing with a subtle digital push-in of a few inches over five seconds | warm window daylight in a believable home office; preserve the Avatar's natural low-contrast facial variation, stable realistic skin texture, integrated eye and brow detail, and soft non-oily highlights | quiet room tone and a soft chair creak under the opening lean. Frame: 9:16. Execute AD as one 5-second clip. - Render AD's ordered action, cast, hand/tool choreography, camera intent, cause/effect, dialogue, and end state. Do not add unrelated people, hands, props, scenes, claims, or beats. - Use each attached picture only for its AD REFERENCE PLAN role. Raw product pictures are authoritative for visible packaging, Logo, and exterior form; do not fabricate or morph them. - Do not show any model, person, face, body, or hands unless AD explicitly requires them. If AD has a HAND PLAN, render only its planned-action hands. Without one, do not add product-handling or focal hand action; passive hands of an AD-required visible person may remain natural, non-interacting, and away from the product. - Unless explicitly required by the AD or an attached reference image, no duplicate faces, bodies, twins, or body doubles. - Execute the speaker-subject bindings declared in AD exactly. For an off-screen speaker, retain the named `<Subject N>` and `(Sx)` from AD as the sole audible source; do not create an anonymous narrator, voice actor, or new speaker. Anonymous product-handling hands are silent unless AD explicitly assigns them a speaker ID. Preserve each speaker's exact AD-stated voice description. - DIALOGUE LOCK: For a non-silent clip, one or more `<d>[Language] ...</d>` blocks in source order are the complete final transcript and the only source of spoken words. `[Language]` is control metadata and must not be spoken. Speak every remaining word exactly once, in exactly that order. Do not paraphrase, translate, add, omit, repeat, abbreviate, or substitute any word, and do not use `<cutoff>` for a complete clip. Follow on-camera lip sync versus off-screen speech exactly - Clean footage: no added captions, subtitles, dialogue text, titles, labels, callouts, or overlays. Dialogue is audio-only. Do not visualize added on-screen text. Preserve only text printed on referenced product/packaging. No BGM. - Preserve AD timing, shot order, lighting, identity, product geometry, markings, and quantity.

AD: 5S CLIP PLAN: REFERENCE PLAN: <Picture 1> is the sequence-wide visual identity reference for <Subject 1> (S1), a late-20s Chinese female kitchen creator; she is visible on camera and is the sole audible speaker. <Picture 2> is the authoritative raw product reference for the visible 巧农婆® 莲子百合山药粉 package, logo, and exterior form; preserve it exactly. Product continuity: Keep the yellow patterned 巧农婆® 莲子百合山药粉 pouch upright, sealed, and visually identical to the raw product reference. [Shot 1] 00:00.000-00:05.000 | medium close-up from a tripod-mounted phone at the kitchen-counter height, facing <Subject 1> (S1) one meter away; the sealed pouch stands upright on the wooden counter in the lower foreground, logo facing camera | <Subject 1> begins with a curious smile, looks from the pouch to camera, then makes one small open-palm gesture beside it without touching; the pouch remains stable and legible, and she ends looking directly into camera | dialogue: <d>[Chinese] 想找一款日常冲着喝的谷物粉，我先看中了这袋莲子百合山药粉。</d> | <Subject 1> (S1), on-camera lip sync, young adult female voice, natural Mainland Chinese Mandarin, warm credible creator tone, natural steady conversational pace; same speaking rate across all clips; no rushed delivery, no elongated syllables, no dramatic slowdown. | locked tripod framing with a subtle digital push-in of less than five percent | warm window daylight, lived-in home kitchen, realistic temporally stable natural skin texture and integrated eye, brow, lash, and soft highlight detail; no beauty filter | quiet kitchen room tone and a soft fabric rustle synchronized with the gesture. Frame: 9:16. Execute AD as one 5-second clip. - Render AD's ordered action, cast, hand/tool choreography, camera intent, cause/effect, dialogue, and end state. Do not add unrelated people, hands, props, scenes, claims, or beats. - Use each attached picture only for its AD REFERENCE PLAN role. Raw product pictures are authoritative for visible packaging, Logo, and exterior form; do not fabricate or morph them. - Do not show any model, person, face, body, or hands unless AD explicitly requires them. If AD has a HAND PLAN, render only its planned-action hands. Without one, do not add product-handling or focal hand action; passive hands of an AD-required visible person may remain natural, non-interacting, and away from the product. - Unless explicitly required by the AD or an attached reference image, no duplicate faces, bodies, twins, or body doubles. - Execute the speaker-subject bindings declared in AD exactly. For an off-screen speaker, retain the named `<Subject N>` and `(Sx)` from AD as the sole audible source; do not create an anonymous narrator, voice actor, or new speaker. Anonymous product-handling hands are silent unless AD explicitly assigns them a speaker ID. Preserve each speaker's exact AD-stated voice description. - DIALOGUE LOCK: For a non-silent clip, one or more `<d>[Language] ...</d>` blocks in source order are the complete final transcript and the only source of spoken words. `[Language]` is control metadata and must not be spoken. Speak every remaining word exactly once, in exactly that order. Do not paraphrase, translate, add, omit, repeat, abbreviate, or substitute any word, and do not use `<cutoff>` for a complete clip. Follow on-camera lip sync versus off-screen speech exactly - Clean footage: no added captions, subtitles, dialogue text, titles, labels, callouts, or overlays. Dialogue is audio-only. Do not visualize added on-screen text. Preserve only text printed on referenced product/packaging. No BGM. - Preserve AD timing, shot order, lighting, identity, product geometry, markings, and quantity.

AD: 5S CLIP PLAN: REFERENCE PLAN: <Picture 1> is the supplied Avatar visual identity reference for <Subject 1> (S1), a young adult female presenter; she is not visible in this clip but is the sole audible speaker. <Picture 2> is the authoritative raw product reference for <Subject 2>, the Instant Pot Duo stainless-steel multicooker; preserve its exact black lid, front control panel, Instant Pot Duo logo, and exterior form. <Subject 1> (S1), the same presenter defined by <Picture 1>, is not visible but is the sole audible speaker; no anonymous narrator, voice actor, or handling hand speaks. Product continuity: Keep <Subject 2> upright on a steady kitchen counter with its black lid closed and its exact front panel facing camera. [Shot 1] 0:00.000-0:02.300 | sharp close product shot, tripod-mounted camera at counter height | <Subject 2> stands centered on a clean lived-in kitchen counter, black lid closed, stainless body and front panel clear | no hands enter; the camera makes a small downward reframe from lid to the front controls, retaining the exact exterior form; end with the cooker centered and front panel sharp | dialogue: <d>[English] Pressure cook, slow cook, rice, steam, sauté, yogurt, sterilize, then keep meals warm.</d> | <Subject 1> (S1), off-screen speech by the same presenter, young adult female voice, subtle warm Midwest American accent, clear practical enthusiasm; natural steady conversational pace; same speaking rate across all clips; no rushed delivery, no elongated syllables, no dramatic slowdown. | slow controlled tripod reframe, no shake | soft kitchen window light with realistic stainless reflections; presenter excluded | gentle music hit and faint room tone. [Shot 2] At 0:02.300 | tighter locked close-up on the Instant Pot Duo logo and front control panel | starting with the logo and labeled controls visible on <Subject 2> | rack focus settles from the logo to the control panel; product stays still, closed, upright, and supported on the counter; end on a clean centered product hero composition | dialogue: <d>[English] That’s why it keeps finding its way into family meal plans.</d> | <Subject 1> (S1), off-screen speech by the same presenter, young adult female voice, subtle warm Midwest American accent, clear practical enthusiasm; natural steady conversational pace; same speaking rate across all clips; no rushed delivery, no elongated syllables, no dramatic slowdown. | locked camera, subtle focus pull only | soft kitchen light, presenter excluded | music resolves with a soft button-like chime. Frame: 9:16. Execute AD as one 5-second clip. - Render AD's ordered action, cast, hand/tool choreography, camera intent, cause/effect, dialogue, and end state. Do not add unrelated people, hands, props, scenes, claims, or beats. - Use each attached picture only for its AD REFERENCE PLAN role. Raw product pictures are authoritative for visible packaging, Logo, and exterior form; do not fabricate or morph them. - Do not show any model, person, face, body, or hands unless AD explicitly requires them. If AD has a HAND PLAN, render only its planned-action hands. Without one, do not add product-handling or focal hand action; passive hands of an AD-required visible person may remain natural, non-interacting, and away from the product. - Unless explicitly required by the AD or an attached reference image, no duplicate faces, bodies, twins, or body doubles. - Execute the speaker-subject bindings declared in AD exactly. For an off-screen speaker, retain the named `<Subject N>` and `(Sx)` from AD as the sole audible source; do not create an anonymous narrator, voice actor, or new speaker. Anonymous product-handling hands are silent unless AD explicitly assigns them a speaker ID. Preserve each speaker's exact AD-stated voice description. - DIALOGUE LOCK: For a non-silent clip, one or more `<d>[Language] ...</d>` blocks in source order are the complete final transcript and the only source of spoken words. `[Language]` is control metadata and must not be spoken. Speak every remaining word exactly once, in exactly that order. Do not paraphrase, translate, add, omit, repeat, abbreviate, or substitute any word, and do not use `<cutoff>` for a complete clip. Follow on-camera lip sync versus off-screen speech exactly - Clean footage: no added captions, subtitles, dialogue text, titles, labels, callouts, or overlays. Dialogue is audio-only. Do not visualize added on-screen text. Preserve only text printed on referenced product/packaging. No BGM. - Preserve AD timing, shot order, lighting, identity, product geometry, markings, and quantity.

AD: 5S CLIP PLAN: REFERENCE PLAN: <Picture 1> is the visual identity reference for <Subject 1> (S1), the same recurring presenter; S1 is not visible and is the sole audible speaker in this clip. Exclude S1's face, body, wardrobe, and reference environment from appearing. <Picture 2> is the authoritative raw product reference for <Subject 2>, the SUPOR rice cooker; preserve its silver rounded-square body, black lid, front control panel, and visible SUPOR logo exactly. Product continuity: The SUPOR rice cooker remains closed, upright, and centered on a clean kitchen counter, with its black lid, silver body, front panel, and SUPOR logo visible. [Shot 1] 00:00.000-00:02.400 | close three-quarter hero view from a tripod-mounted phone, slightly above the counter | <Subject 2> begins upright and closed against a softly blurred kitchen backdrop, logo and front panel facing camera | a slow camera slide reveals the rounded silver side and black lid while the cooker remains completely still | dialogue: <d>[English] Its simple design keeps the counter looking tidy while putting the everyday controls upfront.</d> | <Subject 1> (S1), off-screen speech by the same presenter, early-thirties female voice, neutral General American English with a warm Midwest-leaning practical cadence, moderate energy; natural steady conversational pace; same speaking rate across all clips; no rushed delivery, no elongated syllables, no dramatic slowdown. | slow six-inch leftward slider move, even speed | soft natural window light emphasizes the silver finish without changing product geometry | quiet room tone. [Shot 2] At 00:02.400 | frontal close product shot from the same tripod support | <Subject 2> remains upright, closed, and centered with the SUPOR logo and control panel fully unobstructed | the camera settles into a stable hero framing for the final CTA, ending on the unchanged cooker | dialogue: <d>[English] Take a closer look at the product page.</d> | <Subject 1> (S1), off-screen speech by the same presenter, early-thirties female voice, neutral General American English with a warm Midwest-leaning practical cadence, moderate energy; natural steady conversational pace; same speaking rate across all clips; no rushed delivery, no elongated syllables, no dramatic slowdown. | camera settles and holds | bright, clean countertop environment, no on-screen text | quiet room tone. Execute AD as one 5-second clip. - Render the AD's ordered action, cast, tool or hand choreography, camera intent, visual cause and effect, dialogue, and end state visibly. Do not add unrelated people, hands, props, scenes, claims, or creative beats. - Use every attached picture only for the role assigned in AD's REFERENCE PLAN. Use raw product pictures as the visual reference for visible packaging, Logo, and exterior form. Do not fabricate or morph those exterior identity facts. - Do not show any model, person, face, body, or hands unless AD explicitly requires them. If AD has a HAND PLAN, render only its planned-action hands. Without one, do not add product-handling or focal hand action; passive hands of an AD-required visible person may remain natural, non-interacting, and away from the product. - Execute the speaker-subject bindings declared in AD exactly. For an off-screen speaker, retain the named `<Subject N>` and `(Sx)` from AD as the sole audible source; do not create an anonymous narrator, voice actor, or new speaker. Anonymous product-handling hands are silent unless AD explicitly assigns them a speaker ID. Preserve each speaker's exact AD-stated voice description. - DIALOGUE LOCK: For a non-silent clip, one or more `<d>[Language] ...</d>` blocks in source order are the complete final transcript and the only source of spoken words. `[Language]` is control metadata and must not be spoken. Speak every remaining word exactly once, in exactly that order. Do not paraphrase, translate, add, omit, repeat, abbreviate, or substitute any word, and do not use `<cutoff>` for a complete clip. Follow on-camera lip sync versus off-screen speech exactly - Generate clean footage for later subtitle compositing: no added captions, subtitles, dialogue text, titles, labels, callouts, or text overlays in any frame. Dialogue is audio-only. If AD mentions added on-screen text, do not visualize it. Preserve only text already printed on the referenced product or packaging. No BGM. - Preserve AD timing, shot order, cast, actions, camera, lighting, identity, product geometry, markings, quantity, and end state. Add no extra people, hands, props, scenes, or claims.

AD: 5S CLIP PLAN: REFERENCE PLAN: <Picture 1> is the canonical visual identity reference for <Subject 1> (S1), a young adult female pet-owner presenter, visible on camera and the sole audible speaker. <Picture 2> is the authoritative raw product reference for the Greenies Original Regular bag packaging, Logo, and chew exterior form. <Subject 2> is the Greenies bag. <Subject 3> is one chew, matching the raw product reference. HAND PLAN: <Subject 1> (S1) right hand, lower-center frame, starts resting on the countertop beside <Subject 2>, contacts the bag's upper side panel, then makes one small steadying touch and ends beside the bag; the countertop supports <Subject 2> throughout. Product continuity: Preserve the green-and-white Greenies Original Regular bag and its dog-weight labeling exactly from the raw product reference; one green textured chew remains beside the upright bag. [Shot 1] 00:00.000-00:05.000 | medium close-up from a phone on a locked countertop tripod at eye height | <Subject 1> stands behind the counter, facing camera; <Subject 2> is upright front-facing at lower center and <Subject 3> rests beside it | <Subject 1> makes the single planned right-hand steadying touch beside the bag, then looks from the bag to camera with a candid smile; bag stays upright, chew stays still, and the final frame keeps both product items fully legible | dialogue: <d>[English] I found a daily dental chew sized for dogs between twenty-five and fifty pounds.</d> | <Subject 1> (S1), on-camera lip sync, young adult female voice, neutral contemporary General American English, warm practical pet-owner tone, moderate energy; natural steady conversational pace; same speaking rate across all clips; no rushed delivery, no elongated syllables, no dramatic slowdown. | locked tripod camera with a subtle two-inch digital push-in over five seconds | bright natural kitchen window light; temporally stable realistic skin texture, natural eye and lash detail, soft controlled non-oily highlights | faint bag rustle at the touch and quiet room tone. Frame: 9:16. Execute AD as one 5-second clip. - Render AD's ordered action, cast, hand/tool choreography, camera intent, cause/effect, dialogue, and end state. Do not add unrelated people, hands, props, scenes, claims, or beats. - Use each attached picture only for its AD REFERENCE PLAN role. Raw product pictures are authoritative for visible packaging, Logo, and exterior form; do not fabricate or morph them. - Do not show any model, person, face, body, or hands unless AD explicitly requires them. If AD has a HAND PLAN, render only its planned-action hands. Without one, do not add product-handling or focal hand action; passive hands of an AD-required visible person may remain natural, non-interacting, and away from the product. - Unless explicitly required by the AD or an attached reference image, no duplicate faces, bodies, twins, or body doubles. - Execute the speaker-subject bindings declared in AD exactly. For an off-screen speaker, retain the named `<Subject N>` and `(Sx)` from AD as the sole audible source; do not create an anonymous narrator, voice actor, or new speaker. Anonymous product-handling hands are silent unless AD explicitly assigns them a speaker ID. Preserve each speaker's exact AD-stated voice description. - DIALOGUE LOCK: For a non-silent clip, one or more `<d>[Language] ...</d>` blocks in source order are the complete final transcript and the only source of spoken words. `[Language]` is control metadata and must not be spoken. Speak every remaining word exactly once, in exactly that order. Do not paraphrase, translate, add, omit, repeat, abbreviate, or substitute any word, and do not use `<cutoff>` for a complete clip. Follow on-camera lip sync versus off-screen speech exactly - Clean footage: no added captions, subtitles, dialogue text, titles, labels, callouts, or overlays. Dialogue is audio-only. Do not visualize added on-screen text. Preserve only text printed on referenced product/packaging. No BGM. - Preserve AD timing, shot order, lighting, identity, product geometry, markings, and quantity.

AD: 5S CLIP PLAN: REFERENCE PLAN: <Picture 1> is the visual identity reference for recurring <Subject 1> (S1), the same presenter defined by that picture; <Subject 1> (S1) is not visible, and is the sole audible speaker in this clip. <Picture 2> is the authoritative raw product reference for the visible 巧农婆® 莲子百合山药粉 package, logo, and exterior form; preserve it exactly. HAND PLAN: Anonymous handler right hand, lower-right frame, starts holding a plain wooden spoon above a clear glass on the counter; spoon bowl contacts the water surface once, stirs clockwise in one small circle, then ends above the glass. The glass stays supported by the wooden counter; no other hands or people appear. Product continuity: Use the exact yellow patterned pouch from the raw product reference, upright behind a clear glass of light beige prepared drink; do not alter packaging or logo. [Shot 1] 00:00.000-00:02.200 | close product shot from a tripod-mounted phone, slightly above counter level; the exact pouch is upright behind a clear glass already filled with light beige drink, spoon held above the glass | the anonymous right hand lowers the spoon to the liquid surface and makes one slow clockwise stir; the drink forms a small smooth swirl while pouch and glass stay supported and still | dialogue: <d>[Chinese] 包装上写着甄选地道食材，莲子、百合和山药搭配得很清楚。</d> | <Subject 1> (S1), off-screen speech by the same presenter, young adult female voice, natural Mainland Chinese Mandarin, warm credible creator tone, natural steady conversational pace; same speaking rate across all clips; no rushed delivery, no elongated syllables, no dramatic slowdown. | locked close-up, no camera movement | soft daylight across the wood counter, clean realistic kitchen setting | spoon lightly touches glass and liquid swirls softly. [Shot 2] At 00:02.200 | close product shot from the same tripod position; the spoon has stopped above the glass, the swirl settles, and the pouch remains centered behind it | the hand exits downward-right without touching the pouch; the drink surface becomes still and the package remains upright as the stable end state | dialogue: none | no speaker | locked framing, no motion | soft daylight and shallow natural depth of field | the final faint spoon clink fades into quiet room tone. Frame: 9:16. Execute AD as one 5-second clip. - Render AD's ordered action, cast, hand/tool choreography, camera intent, cause/effect, dialogue, and end state. Do not add unrelated people, hands, props, scenes, claims, or beats. - Use each attached picture only for its AD REFERENCE PLAN role. Raw product pictures are authoritative for visible packaging, Logo, and exterior form; do not fabricate or morph them. - Do not show any model, person, face, body, or hands unless AD explicitly requires them. If AD has a HAND PLAN, render only its planned-action hands. Without one, do not add product-handling or focal hand action; passive hands of an AD-required visible person may remain natural, non-interacting, and away from the product. - Unless explicitly required by the AD or an attached reference image, no duplicate faces, bodies, twins, or body doubles. - Execute the speaker-subject bindings declared in AD exactly. For an off-screen speaker, retain the named `<Subject N>` and `(Sx)` from AD as the sole audible source; do not create an anonymous narrator, voice actor, or new speaker. Anonymous product-handling hands are silent unless AD explicitly assigns them a speaker ID. Preserve each speaker's exact AD-stated voice description. - DIALOGUE LOCK: For a non-silent clip, one or more `<d>[Language] ...</d>` blocks in source order are the complete final transcript and the only source of spoken words. `[Language]` is control metadata and must not be spoken. Speak every remaining word exactly once, in exactly that order. Do not paraphrase, translate, add, omit, repeat, abbreviate, or substitute any word, and do not use `<cutoff>` for a complete clip. Follow on-camera lip sync versus off-screen speech exactly - Clean footage: no added captions, subtitles, dialogue text, titles, labels, callouts, or overlays. Dialogue is audio-only. Do not visualize added on-screen text. Preserve only text printed on referenced product/packaging. No BGM. - Preserve AD timing, shot order, lighting, identity, product geometry, markings, and quantity.

AD: 5S CLIP PLAN: REFERENCE PLAN: <Picture 1> is attached only as the visual identity and speaker reference for recurring <Subject 1> (S1); <Subject 1> (S1), the same presenter defined by this picture, is not visible but is the sole audible speaker in the clip. <Picture 2> is the authoritative raw product reference for <Subject 2>, the LASHCOCA!NE PLUS tube, cap, branding, black wand, and exterior form. No face, body, wardrobe, or environment from <Picture 1> may appear; only one anonymous product-handling hand is visible and does not speak. HAND PLAN: Anonymous handler's right hand, lower frame center, starts gripping <Subject 2>'s silver cap while the tube rests upright on the vanity; it twists the cap counterclockwise, lifts the cap and attached wand vertically, and ends holding the wand above the open tube. The vanity supports the tube throughout. Product continuity: Preserve the raw-reference tube and silver cap exactly; the cap separates once and remains attached to its black wand with fine precision applicator. [Shot 1] 0:00.000-0:05.000 | macro close-up from a locked tripod camera, angled slightly above a clean vanity surface | <Subject 2> starts closed and upright at frame center | the anonymous right hand grips only the silver cap, twists once, then lifts the cap straight up to reveal the black wand and fine applicator; the tube stays supported and motionless on the vanity, ending open with the wand held directly above it | dialogue: <d>[English] One daily swipe along the lash line makes it easy to keep consistent.</d> | <Subject 1> (S1), off-screen speech by the same presenter, young adult female voice, subtle West Coast General American accent, candid beauty-creator tone, moderate energy; natural steady conversational pace; same speaking rate across all clips; no rushed delivery, no elongated syllables, no dramatic slowdown. | locked-off macro framing, no camera movement | soft window light catches the silver cap and clear tube; uncluttered vanity background | a quiet twist, soft cap release, and faint wand lift sound. Frame: 9:16. Execute AD as one 5-second clip. - Render AD's ordered action, cast, hand/tool choreography, camera intent, cause/effect, dialogue, and end state. Do not add unrelated people, hands, props, scenes, claims, or beats. - Use each attached picture only for its AD REFERENCE PLAN role. Raw product pictures are authoritative for visible packaging, Logo, and exterior form; do not fabricate or morph them. - Do not show any model, person, face, body, or hands unless AD explicitly requires them. If AD has a HAND PLAN, render only its planned-action hands. Without one, do not add product-handling or focal hand action; passive hands of an AD-required visible person may remain natural, non-interacting, and away from the product. - Unless explicitly required by the AD or an attached reference image, no duplicate faces, bodies, twins, or body doubles. - Execute the speaker-subject bindings declared in AD exactly. For an off-screen speaker, retain the named `<Subject N>` and `(Sx)` from AD as the sole audible source; do not create an anonymous narrator, voice actor, or new speaker. Anonymous product-handling hands are silent unless AD explicitly assigns them a speaker ID. Preserve each speaker's exact AD-stated voice description. - DIALOGUE LOCK: For a non-silent clip, one or more `<d>[Language] ...</d>` blocks in source order are the complete final transcript and the only source of spoken words. `[Language]` is control metadata and must not be spoken. Speak every remaining word exactly once, in exactly that order. Do not paraphrase, translate, add, omit, repeat, abbreviate, or substitute any word, and do not use `<cutoff>` for a complete clip. Follow on-camera lip sync versus off-screen speech exactly - Clean footage: no added captions, subtitles, dialogue text, titles, labels, callouts, or overlays. Dialogue is audio-only. Do not visualize added on-screen text. Preserve only text printed on referenced product/packaging. No BGM. - Preserve AD timing, shot order, lighting, identity, product geometry, markings, and quantity.

AD: 5S CLIP PLAN: REFERENCE PLAN: <Picture 1> is the visual identity reference for <Subject 1> (S1), excluded: her face, body, wardrobe, and environment do not appear; <Subject 1> (S1), the same presenter defined by this picture, is not visible but is the sole audible speaker. <Picture 2> is authoritative raw product evidence for the yellow device and white magnetic USB cable; preserve exterior form. HAND PLAN: Anonymous handler’s right hand, lower-right frame, starts beside the cable connector; fingertip contacts the device’s bottom charging contacts, guides the connector straight onto them, then ends hovering beside the attached connector; the tabletop supports the device throughout. [Shot 1] 00:00–00:05 | Locked top-down close-up over pale-yellow tabletop | <Subject 2> lies face-up with its charging end nearest the lower frame; the white magnetic connector rests one finger-width away | the anonymous right hand advances the connector in one straight line, magnetic contact seats it flush, then withdraws while device and attached cable remain stable | dialogue: <d>[Chinese] 磁吸式充电接口连接更利落，收纳和使用时也少一份牵绊。</d> | <Subject 1> (S1), off-screen speech by the same presenter, young adult female voice, calm warm Mainland Chinese Mandarin, credible conversational tone; natural steady conversational pace; same speaking rate across all clips; no rushed delivery, no elongated syllables, no dramatic slowdown. | locked overhead camera | clean soft daylight, realistic cable texture | soft magnetic click and faint room tone. Frame: 9:16. Execute AD as one 5-second clip. - Render AD's ordered action, cast, hand/tool choreography, camera intent, cause/effect, dialogue, and end state. Do not add unrelated people, hands, props, scenes, claims, or beats. - Use each attached picture only for its AD REFERENCE PLAN role. Raw product pictures are authoritative for visible packaging, Logo, and exterior form; do not fabricate or morph them. - Do not show any model, person, face, body, or hands unless AD explicitly requires them. If AD has a HAND PLAN, render only its planned-action hands. Without one, do not add product-handling or focal hand action; passive hands of an AD-required visible person may remain natural, non-interacting, and away from the product. - Unless explicitly required by the AD or an attached reference image, no duplicate faces, bodies, twins, or body doubles. - Execute the speaker-subject bindings declared in AD exactly. For an off-screen speaker, retain the named `<Subject N>` and `(Sx)` from AD as the sole audible source; do not create an anonymous narrator, voice actor, or new speaker. Anonymous product-handling hands are silent unless AD explicitly assigns them a speaker ID. Preserve each speaker's exact AD-stated voice description. - DIALOGUE LOCK: For a non-silent clip, one or more `<d>[Language] ...</d>` blocks in source order are the complete final transcript and the only source of spoken words. `[Language]` is control metadata and must not be spoken. Speak every remaining word exactly once, in exactly that order. Do not paraphrase, translate, add, omit, repeat, abbreviate, or substitute any word, and do not use `<cutoff>` for a complete clip. Follow on-camera lip sync versus off-screen speech exactly - Clean footage: no added captions, subtitles, dialogue text, titles, labels, callouts, or overlays. Dialogue is audio-only. Do not visualize added on-screen text. Preserve only text printed on referenced product/packaging. No BGM. - Preserve AD timing, shot order, lighting, identity, product geometry, markings, and quantity.

AD: 5S CLIP PLAN: REFERENCE PLAN: <Picture 1> is the visual identity reference for <Subject 2> (S1), an adult male presenter; he is not visible, and is the sole audible speaker in this clip. <Picture 1> does not define the product, set, wardrobe, or composition. <Picture 2> is the authoritative raw product reference for <Subject 1>, governing its visible black colorway, hood, drawcords, kangaroo pocket, pocket patch label, silhouette, and material appearance. The referenced model, face, body, wardrobe, and environment are excluded from the frame. Product continuity: <Subject 1> remains an upright black hooded work sweatshirt on a minimal dark stand; preserve its evidenced hood, drawcords, kangaroo pocket, and naturally distant pocket patch label. [Shot 1] | 0:00.000–0:05.000 | centered medium-wide product shot from a tripod-mounted cinema camera at chest height, eight feet in front of the stand | <Subject 1> begins upright and perfectly still on the minimal dark stand, centered in a dark studio beneath one enormous warm volumetric god-ray descending from high above; its black fabric holds deep soft shadows, with the pocket patch label visible only at natural viewing distance | no limbs or hands appear; the sweatshirt, stand, and camera remain fixed while fine dust motes drift slowly downward through the light beam; the shot ends with the same stable upright product state | dialogue: <d>[English] A clean black work hoodie, held still in one beam of light.</d> | <Subject 2> (S1), off-screen speech by the same presenter, young adult male voice, neutral General American English with a subtle practical Midwest cadence, calm premium tone; natural steady conversational pace; same speaking rate across all clips; no rushed delivery, no elongated syllables, no dramatic slowdown. | camera makes an extremely slow, straight push-in of less than one foot | dark premium studio, warm overhead volumetric light, deep soft shadows, no on-screen text or subtitles | near-silent room tone, faint air hush, soft warm piano at slow tempo; no product movement sounds. Execute AD as one 5-second clip. - Render the AD's ordered action, cast, tool or hand choreography, camera intent, visual cause and effect, dialogue, and end state visibly. Do not add unrelated people, hands, props, scenes, claims, or creative beats. - Use every attached picture only for the role assigned in AD's REFERENCE PLAN. Use raw product pictures as the visual reference for visible packaging, Logo, and exterior form. Do not fabricate or morph those exterior identity facts. - Do not show any model, person, face, body, or hands unless AD explicitly requires them. If AD has a HAND PLAN, render only its planned-action hands. Without one, do not add product-handling or focal hand action; passive hands of an AD-required visible person may remain natural, non-interacting, and away from the product. - Execute the speaker-subject bindings declared in AD exactly. For an off-screen speaker, retain the named `<Subject N>` and `(Sx)` from AD as the sole audible source; do not create an anonymous narrator, voice actor, or new speaker. Anonymous product-handling hands are silent unless AD explicitly assigns them a speaker ID. Preserve each speaker's exact AD-stated voice description. - DIALOGUE LOCK: For a non-silent clip, one or more `<d>[Language] ...</d>` blocks in source order are the complete final transcript and the only source of spoken words. `[Language]` is control metadata and must not be spoken. Speak every remaining word exactly once, in exactly that order. Do not paraphrase, translate, add, omit, repeat, abbreviate, or substitute any word, and do not use `<cutoff>` for a complete clip. Follow on-camera lip sync versus off-screen speech exactly - Generate clean footage for later subtitle compositing: no added captions, subtitles, dialogue text, titles, labels, callouts, or text overlays in any frame. Dialogue is audio-only. If AD mentions added on-screen text, do not visualize it. Preserve only text already printed on the referenced product or packaging. No BGM. - Preserve AD timing, shot order, cast, actions, camera, lighting, identity, product geometry, markings, quantity, and end state. Add no extra people, hands, props, scenes, or claims.

AD: 5S CLIP PLAN: REFERENCE PLAN: <Picture 1> is the visual identity reference for <Subject 2> (S1), an adult male presenter; he is not visible, and is the sole audible speaker in this clip. <Picture 1> does not define the product, set, wardrobe, or composition. <Picture 2> is the authoritative raw product reference for <Subject 1>, governing its visible black colorway, hood, drawcords, kangaroo pocket, pocket patch label, silhouette, and material appearance. The referenced model, face, body, wardrobe, and environment are excluded from the frame. Product continuity: <Subject 1> remains upright, centered, and motionless on the same minimal dark stand; retain the naturally distant pocket patch label and the deep-shadowed black fabric. [Shot 1] | 0:00.000–0:05.000 | centered medium product shot from the same tripod-mounted cinema camera, now slightly closer and still at chest height | <Subject 1> begins perfectly still within the single overhead god-ray, with the hood opening, drawcords, and kangaroo pocket clearly legible; the stand stays hidden in shadow below | no limbs or hands appear; the garment does not shift or animate; dust motes continue their slow natural drift through the beam while the camera reveals more of the fabric’s soft shadow contour; the stable end state is the same upright product, never in close-up | dialogue: <d>[English] Deep shadows bring out the hood, drawcords, and roomy front pocket naturally.</d> | <Subject 2> (S1), off-screen speech by the same presenter, young adult male voice, neutral General American English with a subtle practical Midwest cadence, calm premium tone; natural steady conversational pace; same speaking rate across all clips; no rushed delivery, no elongated syllables, no dramatic slowdown. | camera continues one nearly imperceptible straight push-in, no pan or tilt | dark premium studio, one warm volumetric overhead beam, rich black background, no on-screen text or subtitles | faint air hush, occasional delicate dust ambience, soft warm piano continuing slowly; no garment movement sounds. Execute AD as one 5-second clip. - Render the AD's ordered action, cast, tool or hand choreography, camera intent, visual cause and effect, dialogue, and end state visibly. Do not add unrelated people, hands, props, scenes, claims, or creative beats. - Use every attached picture only for the role assigned in AD's REFERENCE PLAN. Use raw product pictures as the visual reference for visible packaging, Logo, and exterior form. Do not fabricate or morph those exterior identity facts. - Do not show any model, person, face, body, or hands unless AD explicitly requires them. If AD has a HAND PLAN, render only its planned-action hands. Without one, do not add product-handling or focal hand action; passive hands of an AD-required visible person may remain natural, non-interacting, and away from the product. - Execute the speaker-subject bindings declared in AD exactly. For an off-screen speaker, retain the named `<Subject N>` and `(Sx)` from AD as the sole audible source; do not create an anonymous narrator, voice actor, or new speaker. Anonymous product-handling hands are silent unless AD explicitly assigns them a speaker ID. Preserve each speaker's exact AD-stated voice description. - DIALOGUE LOCK: For a non-silent clip, one or more `<d>[Language] ...</d>` blocks in source order are the complete final transcript and the only source of spoken words. `[Language]` is control metadata and must not be spoken. Speak every remaining word exactly once, in exactly that order. Do not paraphrase, translate, add, omit, repeat, abbreviate, or substitute any word, and do not use `<cutoff>` for a complete clip. Follow on-camera lip sync versus off-screen speech exactly - Generate clean footage for later subtitle compositing: no added captions, subtitles, dialogue text, titles, labels, callouts, or text overlays in any frame. Dialogue is audio-only. If AD mentions added on-screen text, do not visualize it. Preserve only text already printed on the referenced product or packaging. No BGM. - Preserve AD timing, shot order, cast, actions, camera, lighting, identity, product geometry, markings, quantity, and end state. Add no extra people, hands, props, scenes, or claims.

AD: 5S CLIP PLAN: REFERENCE PLAN: <Picture 1> is the visual identity reference for <Subject 2> (S1), an adult male presenter; he is not visible, and is the sole audible speaker in this clip. <Picture 1> does not define the product, set, wardrobe, or composition. <Picture 2> is the authoritative raw product reference for <Subject 1>, governing its visible black colorway, hood, drawcords, kangaroo pocket, pocket patch label, silhouette, and material appearance. The referenced model, face, body, wardrobe, and environment are excluded from the frame. Product continuity: <Subject 1> stays upright and motionless on the dark stand; preserve the evidenced black hooded silhouette and keep the pocket patch label at natural distance, never close-up. [Shot 1] | 0:00.000–0:05.000 | refined medium-close product framing from the same locked tripod position, with the sweatshirt centered and fully recognizable | <Subject 1> begins unchanged on its stand inside the single immense warm god-ray; black fabric remains softly modeled by deep shadow and the pocket patch label remains small and naturally distant | no limbs or hands appear; nothing on <Subject 1> moves; dust motes float gently through the overhead beam as the last small camera push settles and holds on the complete upright silhouette; end state is a still, centered sweatshirt in the light | dialogue: <d>[English] Paxton hooded work sweatshirt, simple, steady, and ready for your everyday rotation.</d> | <Subject 2> (S1), off-screen speech by the same presenter, young adult male voice, neutral General American English with a subtle practical Midwest cadence, calm premium tone; natural steady conversational pace; same speaking rate across all clips; no rushed delivery, no elongated syllables, no dramatic slowdown. | camera completes a final slow push-in, then holds absolutely steady for the final second | dark premium studio, warm volumetric overhead beam, deep soft shadows, no on-screen text or subtitles | faint air hush and slow soft warm piano resolve gently; no product movement sounds. Execute AD as one 5-second clip. - Render the AD's ordered action, cast, tool or hand choreography, camera intent, visual cause and effect, dialogue, and end state visibly. Do not add unrelated people, hands, props, scenes, claims, or creative beats. - Use every attached picture only for the role assigned in AD's REFERENCE PLAN. Use raw product pictures as the visual reference for visible packaging, Logo, and exterior form. Do not fabricate or morph those exterior identity facts. - Do not show any model, person, face, body, or hands unless AD explicitly requires them. If AD has a HAND PLAN, render only its planned-action hands. Without one, do not add product-handling or focal hand action; passive hands of an AD-required visible person may remain natural, non-interacting, and away from the product. - Execute the speaker-subject bindings declared in AD exactly. For an off-screen speaker, retain the named `<Subject N>` and `(Sx)` from AD as the sole audible source; do not create an anonymous narrator, voice actor, or new speaker. Anonymous product-handling hands are silent unless AD explicitly assigns them a speaker ID. Preserve each speaker's exact AD-stated voice description. - DIALOGUE LOCK: For a non-silent clip, one or more `<d>[Language] ...</d>` blocks in source order are the complete final transcript and the only source of spoken words. `[Language]` is control metadata and must not be spoken. Speak every remaining word exactly once, in exactly that order. Do not paraphrase, translate, add, omit, repeat, abbreviate, or substitute any word, and do not use `<cutoff>` for a complete clip. Follow on-camera lip sync versus off-screen speech exactly - Generate clean footage for later subtitle compositing: no added captions, subtitles, dialogue text, titles, labels, callouts, or text overlays in any frame. Dialogue is audio-only. If AD mentions added on-screen text, do not visualize it. Preserve only text already printed on the referenced product or packaging. No BGM. - Preserve AD timing, shot order, cast, actions, camera, lighting, identity, product geometry, markings, quantity, and end state. Add no extra people, hands, props, scenes, or claims.

目标时长47秒，1080p，9:16竖屏。【素材职责】 @image1 只用于锁定粉色自动搅拌杯外观：哑光粉杯身、银色杯口圈、同色手柄、正面白色Simple Joy字样、侧面圆形按键、内部不锈钢胆与白色星形叶轮；全程保持同一版本与结构。 @image2 只用于锁定永和豆浆大包装外观：黄米拼色袋身、左上YON HO永和豆浆标识、左侧“配料表只有大豆”、右侧大字“纯豆浆粉”、下方40g蛋白质与净含量216克(18克/包x12)排版；全程一致。 @image3 只用于锁定永和豆浆独立小包装外观：浅金色袋身、重复金色YON HO永和豆浆标识与斜纹；全程一致。 @image4 只用于锁定女主外观：东亚年轻女性，黑长直发，浅灰圆领卫衣内搭白边，深蓝直筒牛仔裤，白鞋；全程同一人物与服装。 @image5 只用于锁定安检员外观：东亚年轻男性，黑短发，黑色制服套装；全程同一人物与服装。 @image6 只用于锁定路人外观：东亚中老年男性，黑灰短发，深灰Polo衫，卡其长裤，白鞋；全程同一人物与服装。【全局制作设定】全片共3位出镜人物：女主（ @image4 ）承担主要展示与讲解，安检员（ @image5 ）仅在前段安检区出现，路人（ @image6 ）在中后段候车区出现并发声。发声来源明确为女主、安检员、路人三人同步对白，以及女主在清洗特写时的画外旁白；无额外旁白身份。场景分两处：安检区含传送带、灰白安检门、金属台面，明亮均匀室内光；候车区含白色圆柱、金属台面、浅蓝行李箱，同样明亮均匀光。清洗演示插入居家浅色台面与不锈钢水槽两个特写环境，光线明亮柔和。商品始终为 @image1 粉色搅拌杯，搭配 @image2 大包装与 @image3 小包装，不出现其他品牌或款式。【全片共享风格规则】固定机位为主，平视略俯或正面对称构图，浅景深突出人物与商品；切镜以硬切承接动作完成态，不用炫技转场。表演自然生活化，安检段严肃紧张，候车段轻松好奇，女主语气礼貌热情，路人语气朴实惊叹。声音组织保留安检警报音效、电机转动声、水流声等非语音声音，视频全程不需要BGM。画面不生成任何字幕、标题或逐句文字，仅允许商品包装自身印刷文字自然可见。【一致性确认】人物锁定：女主 @image4 、安检员 @image5 、路人 @image6 三人外观服装全程不变，发声来源明确为三人对白及女主画外旁白，无额外旁白。商品锁定： @image1 粉色搅拌杯版本、颜色、按键位置、内部叶轮结构全程一致； @image2 大包装与 @image3 小包装印刷文字与配色全程一致。场景连续性：安检区与候车区布局固定，清洗特写为独立插入环境，返回候车区时人物姿态与商品状态承接前一镜完成态。功能面朝向：杯身正面Simple Joy字样与侧面按键相对镜头与操作者的朝向在各关键动作中明确，内胆星形叶轮仅在开盖或俯视时可见。操作状态：杯盖开合、充电线插拔、叶轮拆装均有明确初始与完成态，不在高风险动作中途跨越23秒拼接边界。画面元素：不生成任何字幕、标题或营销文字叠加，仅商品包装自身印刷自然可见。【0–5秒｜安检放箱触发警报】中景固定机位，女主位于画面右侧，右手单手提起浅蓝行李箱放到视频左侧安检传送带上，女主左手拿着 @image1 粉色搅拌杯，杯内有半杯清水，底部白色搅拌子叶轮高速旋转产生 visible 漩涡。女主左手持杯居中穿过灰白安检门，门顶右侧红灯闪烁并发出“滴滴滴”警报音效，女主笑容凝固停步转头看向左侧。安检员（ @image5 ）从左侧入画，右手前伸做阻拦示意，表情严肃。安检员对白：“女士，请站住，您携带的杯子异常，需要检查。”【5–8秒｜惊讶解释按停】近景过安检员左肩反拍女主正面，女主瞪大眼睛露出惊讶表情看向安检员，随后低头看左手握着的 @image1 杯内旋转水流。女主抬头急切解释，右手食指伸向杯子把手最上方的圆形按键。女主对白：“不好意思，这是买豆浆送的电动搅拌杯，按一下就关上了。”特写切至杯口上方，可以清晰看到杯子里面的水流和杯子把手最上方的按钮，右手食指按一下按键，杯内水流漩涡随即停止，白色搅拌子星形叶轮静止可见。【8–12秒｜沟通水处理转身】中近景正反打，安检员伸手指向杯中水说话，女主恍然大悟微笑点头。安检员对白：“杯子是没问题，您看下水是要喝掉还是倒掉？”女主对白：“那等我一下。”硬切至候车区白色圆柱旁，女主背对安检区走到柱旁。女主右手掀开 @image1 杯盖放到行李箱拉杆旁，左手从牛仔裤口袋掏出 @image3 浅金色小包装豆浆粉。路人（ @image6 ）在左侧低头看手机，随后转头探身看向女主手中杯子与豆粉。【12–16秒｜倒粉搅拌路人搭讪】近景特写固定机位俯拍 @image1 杯口置于金属台面，女主右手持撕开的 @image3 小袋将浅黄豆粉倒入清水中，左手食指按下杯子把手上方按钮开关，伴随轻微电机声，杯子最底下白色搅拌子叶轮高速旋转瞬间将豆粉卷入水下形成乳白色豆浆。硬切至双人同框中近景，路人凑近盯着旋转杯子好奇发问，女主转头看他。路人对白：“小姑娘，你这杯子挺高级啊，能搅……芝麻糊吗？”【16–21秒｜漩涡特写多功能介绍】极特写垂直俯拍 @image1 杯口，乳白色豆浆高速旋转形成深邃中心漩涡。接下来女主说的蛋白粉、燕麦片、芝麻糊杯子里的溶液跟着变化着，女主画外音延续对白：“蛋白粉、燕麦片、芝麻糊通通都能搅，而且这个还能打鸡蛋。”硬切回女主近景，双手捧着装满豆浆的 @image1 杯于胸前，神情自豪面带灿烂笑容看向左侧路人继续介绍。【21–24秒｜惊叹转速追问清洗】近景过女主肩反拍路人正面，路人身体前倾挑眉瞪眼露出惊叹表情。路人对白：“这杯子力气看着是真大。”硬切至侧面双人同框中景，女主手持没有盖子的 @image1 杯自豪回应。女主对白：“一分钟一万转呢。”硬切近景路人皱眉低头看杯，右手轻扶杯身提问。路人对白：“清洗起来很麻烦吧？”【24–28秒｜自清洗演示拆洗结构】特写固定机位俯拍居家浅色台面上 @image1 杯，前半段内壁有豆浆残留且清水高速旋转冲刷，后半段跳切至同一机位杯壁洁净透明清水继续旋转。女主画外旁白：“喝完加点水转一转，自己就洗干净了，里面的搅拌子可拆， 边边角角都能洗到。”硬切至高角度俯拍不锈钢水槽，单手将杯子里最底部的白色星形叶轮搅拌子拿出，在水流下冲洗。AUDIO DIRECTION: @audio1 is the original instrumental background score for THIS segment only. Use its rhythm and musical progression as the background-music reference. Preserve all dialogue, voiceover and sound-effect instructions in the script, generate them with native audio, and keep speech intelligible above this background. Do not transcribe or sing the music, add lyrics, restart the score, or replace spoken lines with music.

目标时长47秒，1080p，9:16竖屏。【素材职责】 @image1 只用于锁定粉色自动搅拌杯外观：哑光粉杯身、银色杯口圈、同色手柄、正面白色Simple Joy字样、侧面圆形按键、内部不锈钢胆与白色星形叶轮；全程保持同一版本与结构。 @image2 只用于锁定永和豆浆大包装外观：黄米拼色袋身、左上YON HO永和豆浆标识、左侧“配料表只有大豆”、右侧大字“纯豆浆粉”、下方40g蛋白质与净含量216克(18克/包x12)排版；全程一致。 @image3 只用于锁定永和豆浆独立小包装外观：浅金色袋身、重复金色YON HO永和豆浆标识与斜纹；全程一致。 @image4 只用于锁定女主外观：东亚年轻女性，黑长直发，浅灰圆领卫衣内搭白边，深蓝直筒牛仔裤，白鞋；全程同一人物与服装。 @image5 只用于锁定安检员外观：东亚年轻男性，黑短发，黑色制服套装；全程同一人物与服装。 @image6 只用于锁定路人外观：东亚中老年男性，黑灰短发，深灰Polo衫，卡其长裤，白鞋；全程同一人物与服装。【全局制作设定】全片共3位出镜人物：女主（ @image4 ）承担主要展示与讲解，安检员（ @image5 ）仅在前段安检区出现，路人（ @image6 ）在中后段候车区出现并发声。发声来源明确为女主、安检员、路人三人同步对白，以及女主在清洗特写时的画外旁白；无额外旁白身份。场景分两处：安检区含传送带、灰白安检门、金属台面，明亮均匀室内光；候车区含白色圆柱、金属台面、浅蓝行李箱，同样明亮均匀光。清洗演示插入居家浅色台面与不锈钢水槽两个特写环境，光线明亮柔和。商品始终为 @image1 粉色搅拌杯，搭配 @image2 大包装与 @image3 小包装，不出现其他品牌或款式。【全片共享风格规则】固定机位为主，平视略俯或正面对称构图，浅景深突出人物与商品；切镜以硬切承接动作完成态，不用炫技转场。表演自然生活化，安检段严肃紧张，候车段轻松好奇，女主语气礼貌热情，路人语气朴实惊叹。声音组织保留安检警报音效、电机转动声、水流声等非语音声音，视频全程不需要BGM。画面不生成任何字幕、标题或逐句文字，仅允许商品包装自身印刷文字自然可见。【一致性确认】人物锁定：女主 @image4 、安检员 @image5 、路人 @image6 三人外观服装全程不变，发声来源明确为三人对白及女主画外旁白，无额外旁白。商品锁定： @image1 粉色搅拌杯版本、颜色、按键位置、内部叶轮结构全程一致； @image2 大包装与 @image3 小包装印刷文字与配色全程一致。场景连续性：安检区与候车区布局固定，清洗特写为独立插入环境，返回候车区时人物姿态与商品状态承接前一镜完成态。功能面朝向：杯身正面Simple Joy字样与侧面按键相对镜头与操作者的朝向在各关键动作中明确，内胆星形叶轮仅在开盖或俯视时可见。操作状态：杯盖开合、充电线插拔、叶轮拆装均有明确初始与完成态，不在高风险动作中途跨越23秒拼接边界。画面元素：不生成任何字幕、标题或营销文字叠加，仅商品包装自身印刷自然可见。【28–34秒｜追问续航演示充电】硬切回候车区近景过女主肩拍路人，路人盯着女主手中杯子里已变空的干净 @image1 杯追问。路人对白：“那不会续航很短吧？”硬切杯子底部画面近景，女主左手握杯子，右手持两个5号电池，装入杯子底部两个凹槽。局部特写快切展示装电池的画面，随后切回女主抬头微笑。女主对白：“大容量电池款，续航拉满，一次能用好久 ，特别方便。”【34–39秒｜索要链接揭晓赠品】近景过女主肩拍路人，路人眉开眼笑低头双手操作黑色手机。路人对白：“链接发我一个呗，我也去买一个。”硬切反打女主近景，摇头摆手解释，右手从身侧提起 @image2 大包装袋展示在胸前，左手仍握@image1 杯。女主对白：“这不是买的，买豆浆粉就送这个搅拌杯，我也没想到这么好用，就是我喝的这个永和豆浆的。”【39–43秒｜端详包装指认配料】中近景双人同框，路人双手捧着 @image2 大袋低头仔细查看包装正面，女主在右侧微笑注视。路人对白：“这配料表还挺干净，只有东北大豆，而且这个蛋白质也高，挺适合我孙女喝的。”硬切至虚化背景下的 @image2 包装局部特写，一只手食指依次点指左侧“配料表只有大豆”文字与右下角“40g”蛋白质圆标。【43–47秒｜全家福促销CTA收尾】硬切回候车区金属台面中近景，台面左中右依次摆放装有豆浆的 @image1 杯、两袋 @image3 小包装、一大袋 @image2 包装；后方路人看手机微笑，女主用手势配合口播逐一指认。女主先伸左手张开四指比划，再依次指向试喝装与搅拌杯。女主对白：“现在拍一到手7大包，还送这个电动搅拌杯，你赶紧去看看吧。“路人回应说：好嘞。””视频在两人微笑互动中直接结束，最后定帧为台面产品全家福与两人同框。AUDIO DIRECTION: @audio1 is the original instrumental background score for THIS segment only. Use its rhythm and musical progression as the background-music reference. Preserve all dialogue, voiceover and sound-effect instructions in the script, generate them with native audio, and keep speech intelligible above this background. Do not transcribe or sing the music, add lyrics, restart the score, or replace spoken lines with music.

CLIP PLAN: 15 seconds, exact 9:16 vertical. Controlled-realism Mainland China home-health ecommerce UGC, clean daylight at a light home table; natural phone-camera stability with slight handheld drift, no plastic skin, robotic motion, BGM, added speech, dialogue subtitles, or invented on-screen text. Only <d> dialogue is spoken; tags/language labels are silent metadata. Preserve all authentic product print.REFERENCE PLAN: <Picture 1> is the visual identity and speaker reference for <Subject 1> (S1), an early-30s Chinese woman presenter; she is visible on camera in Shot 1, with natural stable facial texture and unfiltered skin. <Picture 2> is authoritative raw packaging/Logo reference: AUX 奥克斯 white-and-blue BSX519 box and its pictured white monitor identity. <Picture 3> and <Picture 4> are authoritative raw product/use references for the white angled monitor, green display, blue round button, gray tube and gray upper-arm cuff. Raw product references control all visible product exterior, Logo and packaging; do not alter or invent components.HAND PLAN: A1—S1 left hand, lower-left frame, starts resting beside the tabletop monitor and ends steadying the monitor’s left side; contact: monitor side only, sole action: stabilizes it on the table. A2—S1 right hand, lower-right frame, starts above the blue round button and ends lifted beside it; contact: button only, sole action: one downward press. A3—S1 left hand, frame-left at her right upper arm, starts holding the loose gray cuff flap and ends releasing the fastened flap; contact: cuff flap only, sole action: pulls it across and presses it to fix the cuff while her right arm remains relaxed and supported on the tabletop.Product continuity: Keep the white AUX 奥克斯 monitor upright and supported on the light tabletop, its gray tube connected to the gray cuff; maintain the verified white, green, blue and gray appearance.VOICE / DELIVERY: <Subject 1> (S1), young adult female voice, Mandarin Chinese; on-camera lip sync then off-camera voiceover by the same presenter, natural conversational delivery; steady voice identity, unforced rhythm, no announcer polish.0:00.000 [Shot 1] | Medium vertical home-table shot: S1 faces camera, the authentic AUX 奥克斯 box stands behind the supported monitor; A1 and A2 make one clear press on the blue button, then S1 looks from monitor to camera→monitor remains upright with its green display visible | <d>[Chinese] 家里备血压计，我更看重操作别复杂。</d> | on-camera lip sync; quiet button tap.0:03.800 [Shot 2] | Cut to a three-quarter tabletop use view including S1’s torso and right upper arm: monitor remains supported in foreground, tube visibly leads to the already-positioned gray cuff; A3 performs one continuous pull-across→press fastening action, leaving cuff flat and secured | <d>[Chinese] 奥克斯这款臂式血压计，三百六十升级袖带，一拉一贴就能固定。</d> | off-camera voiceover; soft fabric-and-hook-and-loop fastening sound, decisive contact remains visible.0:08.500 [Shot 3] | Stable over-table view: cuff stays secured on S1’s relaxed supported arm, monitor remains connected and legible; green display settles while S1’s face stays naturally visible above the table edge→measurement setup holds steady | <d>[Chinese] 充电款，测量时有语音播报，还能查看记忆记录。</d> | off-camera voiceover; restrained operating chime only, no device words.0:12.300 [Shot 4] | Return to a composed medium shot: S1 sits beside the complete connected setup and authentic box, calmly gestures open-palmed without touching the product→end on the monitor, cuff, box, and S1 all stable in frame | <d>[Chinese] 正规医疗器械认证，活动价七十九元。</d> | on-camera lip sync; finish at a complete sentence and hold one beat.

NECESSARY GUIDANCECreate a 15-second vertical 9:16 UGC product talk with one adult US creator in a bright, tidy bathroom vanity setting. Use only the exact Hismile Tooth Armour bottle from @image_1; preserve its pink bottle, white dropper cap, label, and proportions. Fixed rear smartphone camera on a tripod, never selfie or mirror framing. Keep the lower 22% empty, with no speech captions.R2V Prompt:References: @image_1 is the exact Hismile Tooth Armour Toothpaste Serum bottle and box identity reference, used for accurate packaging, label, and pink serum-bottle appearance.Global direction: Realistic, lightly polished creator-shot bathroom routine in soft daylight, fixed stable phone-camera perspective, natural skin texture and practical vanity reflections. One creator, one product, believable hands and continuous spatial continuity. Clear spoken en-US dialogue, soft bathroom room tone, synchronized cap-twist, dropper, serum, and brushing sounds, with a very subtle upbeat instrumental bed under speech. Use a restrained graphic system of bold clean white sans-serif type with deep charcoal lettering on rounded translucent blush-pink panels, small hot-pink accent line, gentle slide-in and no excessive animation. Render only the specified graphic copy, not the spoken words.Timeline:0.0–2.0s: First frame: the creator stands at the bathroom vanity facing the fixed camera, holding the closed exact pink Hismile bottle label-forward beside a clean toothbrush at chest height, mouth unobstructed. With direct eye contact and natural lip sync, she says, “Want an easy enamel-support step?” The bottle catches practical window light; faint room ambience and subtle music begin.2.0–4.0s: Without changing position, she raises the closed bottle closer beside her face, keeping both her mouth and the readable front label visible. She says, “This is Hismile Tooth Armour Toothpaste Serum for everyday enamel support.” A soft handling sound accompanies the movement.4.0–9.0s: Direct cut to a tight overhead vanity close-up of hands only. She lays the toothbrush bristles-up on the dry counter, steadies the bottle, twists the white dropper cap open, lifts the dropper straight out, positions it over the bristles, and dispenses a small glossy serum amount onto them. She returns the dropper into the bottle, twists until closed, then places the closed label-forward bottle beside the prepared toothbrush. Audible cap twist, a quiet serum dispense, and bottle set-down; no speaking or lip movement. On-screen graphic — exact text “Use alone or add to toothpaste”; purpose; upper-middle safe screen region; small centered scale and alignment; bold clean white sans-serif typography; deep charcoal letters with hot-pink accent; rounded translucent blush-pink panel with a fine white edge; gentle side-slide entrance; stable hold; soft fade exit; required foreground/background contrast.9.0–12.0s: Wider fixed shot at the same sink: she picks up the prepared toothbrush, turns slightly toward the basin so her mouth is out of frame, and begins a natural brief brushing-routine motion. Brush sounds are soft and realistic. She lowers the brush toward the sink and sets the closed bottle label-forward on the vanity, beside the sink.12.0–15.0s: She turns back to camera with her mouth fully visible, holds the exact closed bottle at chest height with label readable, and gives a small approving nod. She says, “Add it to your routine—brush for two minutes, then expel.” Music and bathroom ambience continue beneath clear lip-synced speech. On-screen graphic — exact text “Brush 2 min • Expel”; purpose; upper-middle safe screen region; small centered scale and alignment; bold clean white sans-serif typography; deep charcoal letters with hot-pink accent; rounded translucent blush-pink panel with a fine white edge; gentle side-slide entrance; stable hold; no exit; required foreground/background contrast.Final state: The creator faces camera naturally with the closed pink Hismile Tooth Armour bottle label-forward at chest height; the prepared toothbrush rests visibly on the vanity near the sink, the “Brush 2 min • Expel” panel remains stable in the upper-middle safe area, and the lower screen stays clear.

NECESSARY GUIDANCECreate a 15-second vertical 9:16 UGC product talk with one adult US creator in a bright, tidy bathroom vanity setting. Use only the exact Hismile Tooth Armour bottle from @image_1; preserve its pink bottle, white dropper cap, label, and proportions. Fixed rear smartphone camera on a tripod, never selfie or mirror framing. Keep the lower 22% empty, with no speech captions.R2V Prompt:References: @image_1 is the exact Hismile Tooth Armour Toothpaste Serum bottle and box identity reference, used for accurate packaging, label, and pink serum-bottle appearance.Global direction: Realistic, lightly polished creator-shot bathroom routine in soft daylight, fixed stable phone-camera perspective, natural skin texture and practical vanity reflections. One creator, one product, believable hands and continuous spatial continuity. Clear spoken en-US dialogue, soft bathroom room tone, synchronized cap-twist, dropper, serum, and brushing sounds, with a very subtle upbeat instrumental bed under speech. Use a restrained graphic system of bold clean white sans-serif type with deep charcoal lettering on rounded translucent blush-pink panels, small hot-pink accent line, gentle slide-in and no excessive animation. Render only the specified graphic copy, not the spoken words.Timeline:0.0–2.0s: First frame: the creator stands at the bathroom vanity facing the fixed camera, holding the closed exact pink Hismile bottle label-forward beside a clean toothbrush at chest height, mouth unobstructed. With direct eye contact and natural lip sync, she says, “Want an easy enamel-support step?” The bottle catches practical window light; faint room ambience and subtle music begin.2.0–4.0s: Without changing position, she raises the closed bottle closer beside her face, keeping both her mouth and the readable front label visible. She says, “This is Hismile Tooth Armour Toothpaste Serum for everyday enamel support.” A soft handling sound accompanies the movement.4.0–9.0s: Direct cut to a tight overhead vanity close-up of hands only. She lays the toothbrush bristles-up on the dry counter, steadies the bottle, twists the white dropper cap open, lifts the dropper straight out, positions it over the bristles, and dispenses a small glossy serum amount onto them. She returns the dropper into the bottle, twists until closed, then places the closed label-forward bottle beside the prepared toothbrush. Audible cap twist, a quiet serum dispense, and bottle set-down; no speaking or lip movement. On-screen graphic — exact text “Use alone or add to toothpaste”; purpose; upper-middle safe screen region; small centered scale and alignment; bold clean white sans-serif typography; deep charcoal letters with hot-pink accent; rounded translucent blush-pink panel with a fine white edge; gentle side-slide entrance; stable hold; soft fade exit; required foreground/background contrast.9.0–12.0s: Wider fixed shot at the same sink: she picks up the prepared toothbrush, turns slightly toward the basin so her mouth is out of frame, and begins a natural brief brushing-routine motion. Brush sounds are soft and realistic. She lowers the brush toward the sink and sets the closed bottle label-forward on the vanity, beside the sink.12.0–15.0s: She turns back to camera with her mouth fully visible, holds the exact closed bottle at chest height with label readable, and gives a small approving nod. She says, “Add it to your routine—brush for two minutes, then expel.” Music and bathroom ambience continue beneath clear lip-synced speech. On-screen graphic — exact text “Brush 2 min • Expel”; purpose; upper-middle safe screen region; small centered scale and alignment; bold clean white sans-serif typography; deep charcoal letters with hot-pink accent; rounded translucent blush-pink panel with a fine white edge; gentle side-slide entrance; stable hold; no exit; required foreground/background contrast.Final state: The creator faces camera naturally with the closed pink Hismile Tooth Armour bottle label-forward at chest height; the prepared toothbrush rests visibly on the vanity near the sink, the “Brush 2 min • Expel” panel remains stable in the upper-middle safe area, and the lower screen stays clear.

REFERENCE MAP — @image_1: pink Yodutry knit hood-and-neck-gaiter silhouette, faux-fur trim, ribbed knit, and front ties only; do not carry over its model, building, or implied comfort result. @image_2: integrated hood-and-neck-gaiter construction and front-tie detail only; do not carry over its text, icons, indoor setting, model, or claims. @image_3: close texture reference for faux-fur trim and ribbed knit only; do not carry over its text, craftsmanship, durability, density, or comfort claims.GLOBAL SETUP — Authentic 9:16 handheld friend-shot UGC at one recognizable outdoor city bus stop on an overcast breezy winter morning. One young adult woman wears a cream coat and dark pants, carries a brown shoulder bag, and uses the same pink Yodutry knit balaclava hood throughout. Keep bus shelter, route post, sidewalk, wind direction, daylight, and bag continuous. Natural handheld medium shots and tactile close-ups; direct cuts, normal-speed action. Natural wind and distant traffic under a subtle light BGM. No speech captions. Render only the specified graphic copy, not the spoken words.0.0–3.0s — Handheld vertical medium shot, upper-center reserved as empty pale sky and shelter glass -> start: she waits below the clear graphic area, hunched against the breeze; her left hand grips her coat collar, loose hair blows across her cheek, and her right hand brushes the hair aside. Her open brown shoulder bag hangs at frame-right with only a small pink knit edge peeking out, not focal -> action: wind visibly lifts her hair again and she tightens her collar grip -> result: clear before-state, hair across cheek and hunched posture remain visible -> direct cut. Graphic 0.0–3.0, upper-center, centered: “POV: the wind hits right when your bus is due.” Plain high-legibility TikTok sans-serif, normal bold, no decorative or italic styling. Audio: wind and distant traffic.3.0–5.5s — Chest-up handheld shot -> start: same woman and bag; her left hand releases the collar and stabilizes the open bag rim while her right hand reaches inside and lifts out the pink knit hood. She transfers the hood to both hands, one hand on each side of its lower knit opening -> action: she raises the opening above her forehead, pulls it downward over her head with both hands, then slides the integrated ribbed knit section down until it visibly surrounds her neck; faux-fur trim frames the hood opening and the front ties hang loose -> result: head and neck are visibly covered by the single pink hood-and-gaiter layer -> direct cut. Spoken VO, natural and brief: “My collar was not doing enough.”5.5–10.5s — One continuous normal-speed tracking shot, camera walking backward on the same sidewalk -> start: hood and knit gaiter are already on; bag rests on her shoulder, hands free at her sides -> action: she walks several steps into the same visible breeze; camera moves from profile to front, then briefly pushes close to the faux-fur edge and ribbed knit at her neck before returning to her face -> result: pink hood remains around her head and knit gaiter remains around her neck while she walks -> direct cut. Graphic 6.2–9.6, mid-screen but clear of face and trim: “Hood + neck gaiter coverage.” Plain medium-weight TikTok sans-serif. Audio: wind, footsteps, soft fabric movement.10.5–13.0s — Closer chest-up handheld shot in the same bus-stop context -> start: she has stopped walking, hood and gaiter remain in place, approaching bus visible behind her -> action: her right fingertips lightly touch the faux-fur-trimmed edge once, then drop as she looks toward the arriving bus -> result: she stands upright, no longer holding her coat collar -> direct cut. Spoken VO: “Now I’ve got one layer for my head and neck.”13.0–15.0s — Medium shot matching the opening bus-stop angle -> start: same woman upright with bag on shoulder, pink hood and neck gaiter on; bus pulls into the background -> action: she turns toward the bus and takes one calm step forward -> result/final frame: her covered head and neck remain clearly visible against the bus-stop setting. Graphic 13.0–14.1, lower-middle: “Covered for the walk to the bus.” Graphic 14.1–15.0, lower-middle: “Add this hood-and-neck-gaiter layer to your cold-weather routine.” Both plain bold TikTok sans-serif, clear of her hands and product.

REFERENCE MAP — @image_1: pink Yodutry knit hood-and-neck-gaiter silhouette, faux-fur trim, ribbed knit, and front ties only; do not carry over its model, building, or implied comfort result. @image_2: integrated hood-and-neck-gaiter construction and front-tie detail only; do not carry over its text, icons, indoor setting, model, or claims. @image_3: close texture reference for faux-fur trim and ribbed knit only; do not carry over its text, craftsmanship, durability, density, or comfort claims.GLOBAL SETUP — Authentic 9:16 handheld friend-shot UGC at one recognizable outdoor city bus stop on an overcast breezy winter morning. One young adult woman wears a cream coat and dark pants, carries a brown shoulder bag, and uses the same pink Yodutry knit balaclava hood throughout. Keep bus shelter, route post, sidewalk, wind direction, daylight, and bag continuous. Natural handheld medium shots and tactile close-ups; direct cuts, normal-speed action. Natural wind and distant traffic under a subtle light BGM. No speech captions. Render only the specified graphic copy, not the spoken words.0.0–3.0s — Handheld vertical medium shot, upper-center reserved as empty pale sky and shelter glass -> start: she waits below the clear graphic area, hunched against the breeze; her left hand grips her coat collar, loose hair blows across her cheek, and her right hand brushes the hair aside. Her open brown shoulder bag hangs at frame-right with only a small pink knit edge peeking out, not focal -> action: wind visibly lifts her hair again and she tightens her collar grip -> result: clear before-state, hair across cheek and hunched posture remain visible -> direct cut. Graphic 0.0–3.0, upper-center, centered: "POV: the wind hits right when your bus is due." Plain high-legibility TikTok sans-serif, normal bold, no decorative or italic styling. Audio: wind and distant traffic.3.0–5.5s — Chest-up handheld shot -> start: same woman and bag; her left hand releases the collar and stabilizes the open bag rim while her right hand reaches inside and lifts out the pink knit hood. She transfers the hood to both hands, one hand on each side of its lower knit opening -> action: she raises the opening above her forehead, pulls it downward over her head with both hands, then slides the integrated ribbed knit section down until it visibly surrounds her neck; faux-fur trim frames the hood opening and the front ties hang loose -> result: head and neck are visibly covered by the single pink hood-and-gaiter layer -> direct cut. Spoken VO, natural and brief: "My collar was not doing enough."5.5–10.5s — One continuous normal-speed tracking shot, camera walking backward on the same sidewalk -> start: hood and knit gaiter are already on; bag rests on her shoulder, hands free at her sides -> action: she walks several steps into the same visible breeze; camera moves from profile to front, then briefly pushes close to the faux-fur edge and ribbed knit at her neck before returning to her face -> result: pink hood remains around her head and knit gaiter remains around her neck while she walks -> direct cut. Graphic 6.2–9.6, mid-screen but clear of face and trim: "Hood + neck gaiter." Plain medium-weight TikTok sans-serif. Audio: wind, footsteps, soft fabric movement.10.5–13.0s — Closer chest-up handheld shot in the same bus-stop context -> start: she has stopped walking, hood and gaiter remain in place, approaching bus visible behind her -> action: her right fingertips lightly touch the faux-fur-trimmed edge once, then drop as she looks toward the arriving bus -> result: she stands upright, no longer holding her coat collar -> direct cut. Spoken VO: "Now I've got one layer for my head and neck."13.0–15.0s — Medium shot matching the opening bus-stop angle -> start: same woman upright with bag on shoulder, pink hood and neck gaiter on; bus pulls into the background -> action: she turns toward the bus and takes one calm step forward -> result/final frame: her covered head and neck remain clearly visible against the bus-stop setting. Graphic 13.0–14.1, lower-middle: "Covered for the walk to the bus." Graphic 14.1–15.0, lower-middle: "Add this hood and gaiter to your cold-weather routine." Both plain bold TikTok sans-serif, clear of her hands and product.

Vertical 9:16 video advertisement, 20 seconds total. Warm, emotional storytelling style. SCENE 1 (0-3.5s): Medium shot, living room. Granddaughter (Avatar @image11 identity: young woman, brown hair, green eyes, smiling) receives a black gift box with red ribbon from her grandmother. Sunlight, cozy atmosphere. Camera static. VOICEOVER: "My shining star, you make the world sparkle."SCENE 2 (3.5-7.0s): DIRECT SHOT SOURCE @image1. Gold star necklace with pink tourmaline center and diamond-like halo on beige felt roll. Exact composition preserved. Slow subtle push-in. No text overlays. VOICEOVER: "I'm proud of you, and I love you to the moon and back."SCENE 3 (7.0-11.0s): Over-the-shoulder close-up. Granddaughter holds the open message card with the pink necklace attached (referencing @image3 layout but pink stone). She reads and smiles, touching the pendant. Identity constraint @image11. Handheld float. VOICEOVER: "This custom sparkle star is a constant reminder of our bond."SCENE 4 (11.0-15.0s): Close-up on neck/collarbone. Granddaughter wearing the pink star necklace. Hands frame the jewelry gently. Skin tone natural, lighting soft. Slow pan up. VOICEOVER: "Wear it always, knowing you're extraordinary."SCENE 5 (15.0-20.0s): DIRECT SHOT SOURCE @image10. Black gift box with red ribbon and open display case with necklace and card on marble surface. Premium presentation. Slow zoom out. VOICEOVER: "Coco Jewelry. Celebrate her journey."AUDIO: BGM ON (soft, acoustic piano/guitar, warm and sentimental). VOICEOVER ON (Female, warm, mature, nurturing tone, en-US). VISUAL TEXT/CALLOUTS: 1. At 4.0s, center screen, elegant serif font, white with soft shadow: "Pink Tourmaline".2. At 8.0s, lower third, sans-serif, clean white: "Personalized Birthstone".3. At 16.0s, bottom center, brand font: "Coco Jewelry".BRAND CONSTRAINTS: Preserve authentic 'Coco Jewelry' mark only where visible in source @image10. Do not invent logos. No speech subtitles generated in-video. Keep lower 22% clear for post-production captions.

Vertical 9:16 video advertisement, 20 seconds total. Warm, emotional storytelling style.SCENE 1 (0-3.5s): Medium shot, living room. Granddaughter (Avatar @image11 identity: young woman, brown hair, green eyes, smiling) receives a black gift box with red ribbon from her grandmother. Sunlight, cozy atmosphere. Camera static. VOICEOVER: "My shining star, you make the world sparkle."SCENE 2 (3.5-7.0s): DIRECT SHOT SOURCE @image1. Gold star necklace with pink tourmaline center and diamond-like halo on beige felt roll. Exact composition preserved. Slow subtle push-in. No text overlays. VOICEOVER: "I'm proud of you, and I love you to the moon and back."SCENE 3 (7.0-11.0s): Over-the-shoulder close-up. Granddaughter holds the open message card with the pink necklace attached (referencing @image3 layout but pink stone). She reads and smiles, touching the pendant. Identity constraint @image11. Handheld float. VOICEOVER: "This custom sparkle star is a constant reminder of our bond."SCENE 4 (11.0-15.0s): Close-up on neck/collarbone. Granddaughter wearing the pink star necklace. Hands frame the jewelry gently. Skin tone natural, lighting soft. Slow pan up. VOICEOVER: "Wear it always, knowing you're extraordinary."SCENE 5 (15.0-20.0s): DIRECT SHOT SOURCE @image10. Black gift box with red ribbon and open display case with necklace and card on marble surface. Premium presentation. Slow zoom out. VOICEOVER: "Coco Jewelry. Celebrate her journey."AUDIO: BGM ON (soft, acoustic piano/guitar, warm and sentimental). VOICEOVER ON (Female, warm, mature, nurturing tone, en-US).VISUAL TEXT/CALLOUTS:1. At 4.0s, center screen, elegant serif font, white with soft shadow: "Pink Stone".2. At 8.0s, lower third, sans-serif, clean white: "Birthstone Gift".3. At 16.0s, bottom center, brand font: "Coco Jewelry".BRAND CONSTRAINTS: Preserve authentic 'Coco Jewelry' mark only where visible in source @image10. Do not invent logos. No speech subtitles generated in-video. Keep lower 22% clear for post-production captions.

Wan 3.0 Prompt: Summary: A feminine blue-and-white floral maxi dress proves its defined waist and airy tiered movement in a sunlit garden-path paid social ad. References: @Image1: Authoritative product reference for the exact white base, blue floral print, sleeveless flutter-cap sleeve silhouette, gathered tie neckline, smocked waist, tiered maxi skirt, and fabric appearance. Shots: Shot 1 [0.0-1.5s]: Wide shot — Subject: an adult woman already wearing the exact @Image1 dress with simple neutral sandals. Action: she takes one confident forward step, immediately showing the full flowing maxi silhouette. Setting and lighting: one sunlit landscaped garden path with softly blurred greenery, warm natural daylight. Camera: low-to-eye-level slow push in. Graphics: Visible copy: "Garden-ready flow" Style: small elegant white serif with subtle blue shadow in open upper-side negative space, inside center 4:5 safe area Animation: appear after the shot begins; hold; disappear before the cut. Shot 2 [1.5-4.5s]: Full-body tracking shot — Subject: same woman and unchanged complete outfit. Action: she walks naturally toward camera and makes a gentle half-turn; the smocked waist remains defined as the tiered skirt swings outward and settles with each step. Setting and lighting: same continuous garden path and daylight direction. Camera: backward tracking shot. Graphics: Visible copy: "Defined waist. Easy sway." Style: refined blue sans serif in clear side space beside, never over, the waist or skirt, inside center 4:5 safe area Animation: appear after the shot begins; hold; disappear before the cut. Shot 3 [4.5-7.5s]: Medium three-quarter side shot to close-up — Subject: same woman in the unchanged dress. Action: she completes a light reachable turn and settles; direct cut closer as one hand lightly touches the neckline tie, revealing the gathered neckline, blue floral print, smocked waist texture, and horizontal tier seams while the fabric falls naturally. Setting and lighting: same garden path, soft sunlit greenery. Camera: slow orbit clockwise, then rack focus to the waist and skirt seam. Graphics: none. Shot 4 [7.5-10.0s]: Wide full-body front shot — Subject: same woman and unchanged outfit. Action: she gently holds one side of the skirt, releases it for one final soft sway, then faces camera in a confident relaxed pose with the entire dress unobstructed. Setting and lighting: same garden path, warm daylight. Camera: static locked. Graphics: Visible copy: "Shop the dress" Style: clean white text in a slim deep-blue rounded button, placed in open lower-side space above the subtitle-clear lower 22%, inside center 4:5 safe area Animation: appear after the shot begins; hold through the final frame. Audio: Voice lines: 1.5-4.5s, female voiceover, warm and conversational, en-US: "A defined waist, with an easy flowing skirt." Sound effects: Soft footsteps on the garden path and faint fabric swish synchronized to the walk, turn, and skirt release; subtle outdoor ambience. Music: Light, breezy instrumental music under the voiceover. Final frame: Full-body front composition in the central 4:5 crop-safe area: model confidently faces camera after the skirt settles, with the gathered tie neckline, defined smocked waist, blue floral print, tiered maxi length, and neutral sandals fully visible. Soft greenery frames the background; face and dress stay unobstructed. "Shop the dress" remains stable in open lower-side negative space above the clear lower 22%.

Wan 3.0 Prompt: Summary: A feminine blue-and-white floral maxi dress proves its defined waist and airy tiered movement in a sunlit garden-path paid social ad. References: @Image1: Authoritative product reference for the exact white base, blue floral print, sleeveless flutter-cap sleeve silhouette, gathered tie neckline, smocked waist, tiered maxi skirt, and fabric appearance. Shots: Shot 1 [0.0-1.5s]: Wide shot — Subject: an adult woman already wearing the exact @Image1 dress with simple neutral sandals. Action: she takes one confident forward step, immediately showing the full flowing maxi silhouette. Setting and lighting: one sunlit landscaped garden path with softly blurred greenery, warm natural daylight. Camera: low-to-eye-level slow push in. Graphics: Visible copy: "Garden Flow" Style: small elegant white serif with subtle blue shadow in open upper-side negative space, inside center 4:5 safe area Animation: appear after the shot begins; hold; disappear before the cut. Shot 2 [1.5-4.5s]: Full-body tracking shot — Subject: same woman and unchanged complete outfit. Action: she walks naturally toward camera and makes a gentle half-turn; the smocked waist remains defined as the tiered skirt swings outward and settles with each step. Setting and lighting: same continuous garden path and daylight direction. Camera: backward tracking shot. Graphics: Visible copy: "Defined Waist. Easy Sway." Style: refined blue sans serif in clear side space beside, never over, the waist or skirt, inside center 4:5 safe area Animation: appear after the shot begins; hold; disappear before the cut. Shot 3 [4.5-7.5s]: Medium three-quarter side shot to close-up — Subject: same woman in the unchanged dress. Action: she completes a light reachable turn and settles; direct cut closer as one hand lightly touches the neckline tie, revealing the gathered neckline, blue floral print, smocked waist texture, and horizontal tier seams while the fabric falls naturally. Setting and lighting: same garden path, soft sunlit greenery. Camera: slow orbit clockwise, then rack focus to the waist and skirt seam. Graphics: none. Shot 4 [7.5-10.0s]: Wide full-body front shot — Subject: same woman and unchanged outfit. Action: she gently holds one side of the skirt, releases it for one final soft sway, then faces camera in a confident relaxed pose with the entire dress unobstructed. Setting and lighting: same garden path, warm daylight. Camera: static locked. Graphics: Visible copy: "Shop Now" Style: clean white text in a slim deep-blue rounded button, placed in open lower-side space above the subtitle-clear lower 22%, inside center 4:5 safe area Animation: appear after the shot begins; hold through the final frame. Audio: Voice lines: 1.5-4.5s, female voiceover, warm and conversational, en-US: "A defined waist, with an easy flowing skirt." Sound effects: Soft footsteps on the garden path and faint fabric swish synchronized to the walk, turn, and skirt release; subtle outdoor ambience. Music: Light, breezy instrumental music under the voiceover. Final frame: Full-body front composition in the central 4:5 crop-safe area: model confidently faces camera after the skirt settles, with the gathered tie neckline, defined smocked waist, blue floral print, tiered maxi length, and neutral sandals fully visible. Soft greenery frames the background; face and dress stay unobstructed. "Shop Now" remains stable in open lower-side negative space above the clear lower 22%.

AD: 5S CLIP PLAN: REFERENCE PLAN: <Picture 1> is the sequence-wide visual identity reference for <Subject 1> (S1), an adult female presenter; her face, body, wardrobe, and reference environment are excluded. It is only the visual identity reference for <Subject 1> (S1), who is not visible but is the sole audible speaker in the clip; no anonymous narrator, voice actor, or handling hand speaks. <Picture 2> is authoritative raw product evidence for <Subject 2>, the gray Lexar JumpDrive D400 128GB dual USB-C and USB-A drive and its packaging; it governs product identity, shape, finish, markings, and connector layout. <Picture 3> is supporting raw product evidence confirming the dual-interface product form and compatible-device context; it may supplement but not override <Picture 2>. Product continuity: <Subject 2> remains one gray metal Lexar JumpDrive D400 128GB drive with the USB-C connector on one end and the USB-A connector on the other. 0–5s [Shot 1] | vertical macro close-up from a locked third-person tabletop camera, lens level with a smartphone USB-C port at frame right | <Subject 2> is already held horizontally by anonymous product-handling hands at frame left, USB-C tip aligned one finger-width from the phone port; the phone lies flat on a wood desk, laptop remains softly blurred behind it | Starting state: anonymous right thumb and index finger grip the metal body while the left hand lightly steadies the phone edge. The right hand moves <Subject 2> straight forward on the port axis, makes visible contact, and seats the USB-C connector fully; the right hand releases, leaving the drive supported by the phone. The left hand then lifts away; stable end state is the drive securely connected, its body and keyring opening visible | dialogue: <d>[English] It’s the Lexar JumpDrive D400: one hundred twenty-eight gigabytes, with USB-C and USB-A.</d> | <Subject 1> (S1), off-screen speech by the same presenter, young adult female voice, neutral General American English, practical upbeat paid-social delivery, moderate pace | locked-off macro framing, no camera movement | natural window light across the desk; no presenter appears on screen | soft connector seat click exactly at insertion, faint room tone. Execute AD as one 5-second clip. - Use every attached picture only for the role assigned in AD's REFERENCE PLAN. Raw product references are authoritative for product identity, packaging, logo/text, color, shape, structure, quantity, orientation, and mechanism. - Do not show any model, person, face, body, or hands unless AD explicitly requires them. - Execute the speaker-subject bindings declared in AD exactly. For an off-screen speaker, retain the named `<Subject N>` and `(Sx)` from AD as the sole audible source; do not create an anonymous narrator, voice actor, or new speaker. Anonymous product-handling hands are silent unless AD explicitly assigns them a speaker ID. Preserve each speaker's exact AD-stated voice description. - Inside `<d>...</d>`, `[Language]` is control metadata and must not be spoken. Speak the remaining words exactly once without paraphrasing, translating, adding, omitting, or repeating. Follow on-camera lip sync versus off-screen speech exactly. - Generate clean footage for later subtitle compositing: no added captions, subtitles, dialogue text, titles, labels, callouts, or text overlays in any frame. Dialogue is audio-only. If AD mentions added on-screen text, do not visualize it. Preserve only text already printed on the referenced product or packaging. No BGM. - Preserve AD timing, shot order, cast, actions, camera, lighting, identity, product geometry, markings, quantity, and end state. Add no extra people, hands, props, scenes, or claims.

AD: 5S CLIP PLAN: REFERENCE PLAN: <Picture 1> is the canonical visual identity and speaker reference for <Subject 1> (S1), the mother; she is visible on camera and is the sole audible speaker. <Picture 2> is authoritative raw product evidence for the SYLVOX outdoor TV’s exterior form, bezel, complete stone-wall installation, and family patio context. <Subject 2> is the father; <Subject 3> is the daughter; <Subject 4> is the son. Product continuity: Keep the entire wall-mounted television and stone-wall installation visible throughout, with a moving baseball game, no UI, no subtitles, no stand, and no visible mounting hardware. [Shot 1] 0:00.000–0:01.500 | wide third-person tripod hero view from behind the lounge at seated eye level | the family sits closely together around the barbecue table beneath the complete wall-mounted TV; <Subject 1> sits beside the children, <Subject 2> sits near <Subject 4>, and food remains stably on the table | the family relaxes and watches the active baseball game together; no product contact occurs; end with their attention on the full screen | dialogue: none | no speaker | locked-off hero composition | golden afternoon light with stable natural facial detail for visible faces | soft conversation, plates settling, distant birds. [Shot 2] At 00:01.500 | wide fixed camera retains family and entire TV in one frame | the wall-mounted TV remains fully visible above the seated family | a runner slides toward home plate on the TV; the family gives one brief seated cheer, remaining safely in their seats; end with smiles and attention returning to the screen | dialogue: none | no speaker | fixed camera | warm late-afternoon light, readable vivid screen | short family cheer and soft broadcast crowd sound. [Shot 3] At 00:03.000 | wide tripod hero view, slow gentle push-in while retaining complete TV and wall | <Subject 1> turns from the game toward camera while remaining seated with her family; <Subject 2>, <Subject 3>, and <Subject 4> stay close beside her | <Subject 1> smiles to camera, then turns back toward the family as they laugh together; end with the complete installation visible above the warm gathering | dialogue: <d>[English] It made our backyard feel useful again.</d> | <Subject 1> (S1), on-camera lip sync, early-forties female voice with warm natural US South family-lifestyle cadence, moderate energy; natural steady conversational pace; same speaking rate across all clips; no rushed delivery, no elongated syllables, no dramatic slowdown. | slow six-inch push-in | preserve <Subject 1> facial identity, realistic temporally stable skin texture and soft controlled highlights | shared laughter, patio ambience, low baseball broadcast. Frame: 9:16. Execute AD as one 5-second clip. - Render AD's ordered action, cast, hand/tool choreography, camera intent, cause/effect, dialogue, and end state. Do not add unrelated people, hands, props, scenes, claims, or beats. - Use each attached picture only for its AD REFERENCE PLAN role. Raw product pictures are authoritative for visible packaging, Logo, and exterior form; do not fabricate or morph them. - Do not show any model, person, face, body, or hands unless AD explicitly requires them. If AD has a HAND PLAN, render only its planned-action hands. Without one, do not add product-handling or focal hand action; passive hands of an AD-required visible person may remain natural, non-interacting, and away from the product. - Unless explicitly required by the AD or an attached reference image, no duplicate faces, bodies, twins, or body doubles. - Execute the speaker-subject bindings declared in AD exactly. For an off-screen speaker, retain the named `<Subject N>` and `(Sx)` from AD as the sole audible source; do not create an anonymous narrator, voice actor, or new speaker. Anonymous product-handling hands are silent unless AD explicitly assigns them a speaker ID. Preserve each speaker's exact AD-stated voice description. - DIALOGUE LOCK: For a non-silent clip, one or more `<d>[Language] ...</d>` blocks in source order are the complete final transcript and the only source of spoken words. `[Language]` is control metadata and must not be spoken. Speak every remaining word exactly once, in exactly that order. Do not paraphrase, translate, add, omit, repeat, abbreviate, or substitute any word, and do not use `<cutoff>` for a complete clip. Follow on-camera lip sync versus off-screen speech exactly - Clean footage: no added captions, subtitles, dialogue text, titles, labels, callouts, or overlays. Dialogue is audio-only. Do not visualize added on-screen text. Preserve only text printed on referenced product/packaging. No BGM. - Preserve AD timing, shot order, lighting, identity, product geometry, markings, and quantity.