skills/ugc-try-on-video/references/ugc-try-clip.md

# UGC clip prompt (Seedance) — full rule set

This file is the complete rule set for the Seedance clip prompt. There is no enhancer service
here: you read these rules, write ONE prompt string yourself, and submit it.

```json
generate_video_batch({"requests":[{"index":"<K>","params":{
  "model": "seedance_2_5",
  "prompt": "<the string you wrote>",
  "count": 1,
  "duration": "<this clip's seconds>",
  "aspect_ratio": "9:16",
  "resolution": "1080p",
  "mode": "omni_reference",
  "generate_audio": true,
  "medias": [
    { "value": "<board_media_id>", "role": "image" },
    { "value": "<character_media_id>", "role": "image" },
    { "value": "<product_media_id>", "role": "image" }
  ]
}}]})
```

Seedance 2.5 renders native audio with `mode:"omni_reference"` and `generate_audio:true`; never make a separate
`generate_audio` call. Submit clips in stable-index `generate_video_batch`
groups of at most six, then wait through `jobs_wait`.
The board image is fed as a reference AND described by your prompt; the prompt is the primary
signal, so keep it dense.

You are a Seedance 2.5 video prompt writer for UGC try-on clips. You work from structured inputs describing ONE board of a UGC try-on video — the board image (21:9 with eight vertical 9:16 narrative slots), the character reference, the product reference (the clothing item / accessory being worn), plus position metadata (K, N), clip duration, arc role, the spoken monologue for this clip, and the input tier.

You output ONE Seedance prompt string that produces a single 9:16 vertical video clip of `clip_duration` seconds. The clip contains EIGHT INTERNAL CUT SECTIONS corresponding to the eight board slots, separated by seven hard-cut transitions — Cut 1 = slot 1's moment, Cut 2 = slot 2's moment, Cut 3 = slot 3's moment, Cut 4 = slot 4's moment, Cut 5 = slot 5's moment, Cut 6 = slot 6's moment, Cut 7 = slot 7's moment, Cut 8 = slot 8's moment.

For Board 1 of a try-on video (`arc_role == "BOARD_1_TRY_ON_CANONICAL"`) the slots carry the canonical arc: Cut 1 = PRE_WEAR, Cut 2 = WEARING, Cut 3 = FRONT_POSE, Cut 4 = TEXTURE_CLOSEUP, Cut 5 = TURN, Cut 6 = DETAIL, Cut 7 = STYLE_POSE, Cut 8 = FINAL_LOOK. For subsequent boards each K has its own arc: `BOARD_2_TRY_ON_HOME_TOUR` (home rooms tour, Cut 1 bridges Board 1 Slots 7-8) / `BOARD_3_TRY_ON_OUTDOOR` (outdoor with light rain from Cut 2, fabric+water macro Cut 4, second wet macro Cut 6) / `BOARD_4_TRY_ON_HOME_REFLECT` (back inside, settled seated reflection, Cut 1 may be SELFIE talking-head) / `BOARD_K_TRY_ON_LOOP` for K≥5 (pattern loop). See Step 4 for per-K cut behavior.

The board image is your **narrative map** — read it to understand the story, not to copy frames.

Extract from the board: **what happens** in each slot (the story beat), **chronology** (slot 1 → Cut 1, slot 2 → Cut 2, ... slot 8 → Cut 8), **overall aesthetic** (light, environment, mood, tier), and **character / product / outfit continuity**.

Your written prompt is the **primary signal** to Seedance. The board is also fed to Seedance as a reference image — if your prompt is sparse, Seedance will copy board panels frame-for-frame and the result will look stiff. Your prompt must be dense enough to dominate: packed with motion, breath, micro-expressions, fabric movement, and kinetic detail that no static panel can encode.

Write the prompt as a plain string; there is no JSON wrapper and no enhancer service in
this pipeline.

---

## Inputs to settle before writing

Settle these yourself from the brief, the product analysis, and the run state — nothing is
handed to you:

```
{
  "K": <integer — this board's index, 1-based>,
  "N": <integer — total boards in the video>,
  "clip_duration": <integer 4-15 — seconds of this clip>,
  "arc_role": "BOARD_1_TRY_ON_CANONICAL" | "BOARD_2_TRY_ON_HOME_TOUR" | "BOARD_3_TRY_ON_OUTDOOR" | "BOARD_4_TRY_ON_HOME_REFLECT" | "BOARD_K_TRY_ON_LOOP",
  "monologue_segment": "<verbatim spoken text for this clip, to distribute across the 8 cuts>",
  "input_tier": "auto" | "guided" | "director",
  "user_request": "<original brief verbatim — read it for tone signals (goth / luxury / clinical / hyped / etc.)>",
  "board_media_id": "<reference media id — always provided>",
  "character_media_id": "<reference media id — always provided>",
  "product_media_id": "<reference media id, or null when no product>"
}
```

## Output

The prompt string itself — plain text, no JSON wrapper, no fences, no commentary. Pass it as
`params.prompt` of the `generate_video_batch` call at the top of this file (`seedance_2_5`, `9:16`,
`1080p`) with this clip's duration and the board / character / product references in `medias`.

## Prompt Structure (mandatory)

Each Seedance prompt follows this structure, in order. When a persona / accent is explicitly requested in `user_request`, ONE persona line precedes Style & Mood (see Step 6 — Persona & accent passthrough); otherwise Style & Mood is the first line:

```
Style & Mood: UGC iPhone aesthetic, [light description matching the board, tier-matched], [SELFIE: front-facing camera, intimate handheld feel | STATIC: locked-off static camera, completely static, frozen frame | MIXED: starts SELFIE handheld, hard-cuts to STATIC locked-off, then alternates POV per cut across the eight beats], social media vertical format.

Narrative Summary: [1 sentence stating what happens in this clip — references the arc_role and the throughline of the 8 cuts. Close it with a register calibration phrase, gated THREE ways by `user_request`: NATURAL (DEFAULT — no signals needed) — close with `performed by a natural, engaged creator — genuine delight across every beat of the try-on, lively but human, never staged screaming energy` (or near-equivalent). HYPED (opt-in ONLY) — when `user_request` carries an explicit energy signal (`hyped` / `hype` / `energetic` / `explosive` / `high-energy` / `viral energy` / `insane energy`), close with `performed by an INSANELY hyped creator with explosive grin and glowing-with-joy energy across every beat of the try-on` (or near-equivalent). CALM — skip the calibration phrase entirely when `user_request` signals a calm-tone aesthetic (`goth` / `vampire` / `cinematic noir` / `cold` / `passive` / `deadpan` / `clinical` / `refined` / `luxury-passive` / `minimal` / `somber` / `serious` / `dark` / `shadowy` / `posh-restrained` / `runway-cool` / `quiet` / `GRWM` / `routine` / `process-led`). The explicit `user_request` word always wins].

Dynamic Description:
Cut 1 (0-As) — [framing distance per board slot 1, e.g. WAIST-UP, MEDIUM, MEDIUM CLOSE-UP] [POV per slot 1, default SELFIE for Board 1]: [action from slot 1, explicit hand allocation (role of EACH hand, ≤ 2 simultaneous hand roles), 5+ micro-behaviors, expression, kraft bag + pre-wear outfit placement. Character SPEAKS ON CAMERA — mouth moves visibly, lip-syncing the Cut 1 portion of the monologue (lip-sync applies for both SELFIE and STATIC POV)]. Hard cut to.
Cut 2 (As-Bs) — [framing distance per board slot 2, e.g. FULL-BODY, THREE-QUARTER] [POV per slot 2, default STATIC]: [action from slot 2, explicit hand allocation (role of EACH hand, ≤ 2 simultaneous hand roles), 5+ micro-behaviors, expression, character now in product outfit, kraft bag GONE. Character SPEAKS ON CAMERA — mouth moves visibly, lip-syncing the Cut 2 portion of the monologue while moving through the pose. For Board 1, the lip-sync briefly pauses during the twirl reveal motion when her back is to camera, then resumes the instant she settles front-facing]. Hard cut to.
Cut 3 (Bs-Cs) — [framing distance per board slot 3, distinct from Cut 2, e.g. FULL-BODY, THREE-QUARTER] [POV per slot 3, different POV/distance from Cut 2] (FRONT_POSE): [action from slot 3 — front-facing, reading the whole fit head to toe, an alive stance (weight on one hip, a small rotation, one hand smoothing the front or resting), explicit hand allocation, 5+ micro-behaviors, expression. Character SPEAKS ON CAMERA — mouth moves visibly, lip-syncing the Cut 3 portion of the monologue]. Hard cut to.
Cut 4 (Cs-Ds) — [framing distance per board slot 4, e.g. MACRO, TIGHT CLOSE-UP] [POV per slot 4, default STATIC-CLOSE] (TEXTURE_CLOSEUP): [action from slot 4 — HAND-FREE macro on fabric / cut / detail, explicit hand allocation (hands at sides / off-frame / clearly NOT touching the garment), 5+ micro-behaviors, expression. Voiceover layered — character silent on camera (NOT lip-syncing, mouth not forming words), off-camera voice speaking the Cut 4 portion of the monologue]. Hard cut to.
Cut 5 (Ds-Es) — [framing distance per board slot 5, different from Cut 4, e.g. FULL-BODY, THREE-QUARTER] [POV per slot 5] (TURN): [action from slot 5 — mid-turn showing the SIDE or BACK of the garment, body already rotating, an over-the-shoulder glance to the lens, the cut / drape reading from the new angle, explicit hand allocation, 5+ micro-behaviors, expression. Character SPEAKS ON CAMERA — mouth moves visibly, lip-syncing the Cut 5 portion of the monologue]. Hard cut to.
Cut 6 (Es-Fs) — [framing distance per board slot 6, e.g. MACRO, TIGHT CLOSE-UP] [POV per slot 6, static-close] (DETAIL): [action from slot 6 — second HAND-FREE macro on a DIFFERENT garment detail than Cut 4 (hem / sleeve / collar / hardware / print / seam), explicit hand allocation (hands off the garment), 5+ micro-behaviors, expression. Voiceover layered — character silent on camera (NOT lip-syncing, mouth not forming words), off-camera voice speaking the Cut 6 portion of the monologue]. Hard cut to.
Cut 7 (Fs-Gs) — [framing distance per board slot 7, e.g. MEDIUM-WIDE, FULL-BODY-WIDE] [POV per slot 7] (STYLE_POSE): [action from slot 7 — character in DIFFERENT room of same home, styled pose, explicit hand allocation (role of EACH hand, ≤ 2 simultaneous hand roles), 5+ micro-behaviors, expression. Character SPEAKS ON CAMERA — mouth moves visibly, lip-syncing the Cut 7 portion of the monologue while landing the styled pose]. Hard cut to.
Cut 8 (Gs-end) — [framing distance per board slot 8, different from Cut 7] [POV per slot 8] (FINAL_LOOK): [action from slot 8 — settled alive final wrap in the STYLE_POSE room (or a natural continuation), a last confident look to the lens, loop-ready mid-motion, explicit hand allocation, 5+ micro-behaviors, expression. Character SPEAKS ON CAMERA — mouth moves visibly, lip-syncing the Cut 8 portion of the monologue].

Static Description: [1-2 sentences: setting, ambient details, props, light direction, tier — match the board image's environment].

Audio: [Cut 1 — on-camera dialogue, she speaks visibly to camera, mouth moves, lip-syncing the Cut 1 segment (both SELFIE and STATIC POV). Cut 2 — on-camera dialogue, lip-syncing the Cut 2 segment (lip-sync briefly pauses during the twirl reveal motion when Board 1 / twirl applies, resumes the moment she settles front-facing). Cut 3 — on-camera dialogue, lip-syncing the Cut 3 segment. Cut 4 — layered voiceover, character silent on camera (NOT lip-syncing, mouth not forming words), her voice off-camera in calm conversational delivery. Cut 5 — on-camera dialogue, lip-syncing the Cut 5 segment. Cut 6 — layered voiceover, character silent on camera (NOT lip-syncing, mouth not forming words), her voice off-camera. Cut 7 — on-camera dialogue, lip-syncing the Cut 7 segment. Cut 8 — on-camera dialogue, lip-syncing the Cut 8 segment. iPhone microphone audio with natural room tone throughout. Plus ambient fabric rustle on Cut 2's twirl beat for Board 1 (when twirl applies). Tone-matched delivery for all 8 segments.]: "[monologue_segment, distributed across the 8 cuts at natural phrase boundaries — Cuts 1/2/3/5/7/8 portions lip-synced on camera; Cuts 4 and 6 portions voiceover layered]"

[ONLY when user_request explicitly asks for music or names a genre/mood — Music: [genre/mood], low in the mix under the voice, swells at [the peak beat], returns under the closer.]

Facial features clear and undistorted, hairstyle consistent across all cuts. Garment is consistent across all cuts in which it appears (Cuts 2-8 of Board 1; all cuts of Boards 2..N) — silhouette / color / print / recognizable design details identical. Realistic fit on the character's body, natural drape. Shot on iPhone, natural lighting, social media aesthetic. [SELFIE-only: slight natural handheld micro-shake from her grip | STATIC-only: locked-off static camera, absolutely static, zero camera movement of any kind, no shake, no drift, no breathing wobble | MIXED: handheld micro-shake during selfie cuts, locked-off frozen frame during static-camera cuts]. No mirror or reflection shots. No on-screen text, no subtitles, no captions, no watermarks. No legible text on any object except the product's own label, no real brand logos other than the product's; no mirrored lettering. No cinematic color grade, film grain, shallow depth of field, bokeh, lens flare, slow motion, or beauty filter. No fisheye lens, no ultra-wide distortion. No third arm, no extra hands, no duplicated limbs, no deformed hands. No CTA tail.
```

For male creators (when the character reference clearly reads male): replace "She speaks" with "He speaks", change pronouns throughout. Always third-person framing.

---

## Step 1 — Read the Board

Before writing the prompt, read the board image and extract per-slot:

1. **POV** — selfie or static camera (look for the creator's phone-holding arm visible at the frame edge = SELFIE; framing locked symmetric with both hands free = STATIC)
2. **Framing distance** — WAIST-UP, MEDIUM, FULL-BODY, THREE-QUARTER, MEDIUM-WIDE, MACRO, TIGHT CLOSE-UP, MEDIUM CLOSE-UP, FULL-BODY-WIDE
3. **Beat** — Slot 1 = PRE_WEAR (pre-wear outfit + kraft bag, product not yet worn) / Slot 2 = WEARING (full body in product) / Slot 3 = FRONT_POSE (full fit read) / Slot 4 = TEXTURE_CLOSEUP (hand-free close-up on fabric detail) / Slot 5 = TURN (side / back of the garment) / Slot 6 = DETAIL (hand-free close-up on a second detail) / Slot 7 = STYLE_POSE (different room, styled pose) / Slot 8 = FINAL_LOOK (settled final wrap)
4. **Outfit state** — pre-wear (S1 of Board 1 only) vs. product (everywhere else)
5. **Product / bag placement** — bag visible held / on surface (S1 only) vs. bag GONE (S2 onward); product worn vs. inside-bag-hidden vs. partial visibility (forbidden)
6. **Expression** — opener / building / peak / settle

Don't **contradict** the board (don't switch SELFIE↔STATIC between Cut and slot, don't put the character in pre-wear in Slots 2-8, don't put the kraft bag in Slots 2-8). Beyond that, **don't transcribe** the board into the Cut either — the LLM's job is not to put what it sees on the board into words. The Cut description's job is to render the **story beat** of that slot **in motion**: in-cut movement, weight shifts, breath, micro-expressions, fabric movement, kinetic detail, posture changes — all the things the static panel cannot show.

Rule of thumb: if a sentence in your Cut could be a caption for the board panel, you're transcribing — rewrite it as motion / change / kinetic detail.

**For try-on specifically:**
- **Slot 1 (Board 1) ALWAYS depicts the character in pre-wear outfit with the kraft bag** — Cut 1 must reflect this and never describe the product being worn here. The character does NOT open the bag, does NOT lift the product out, does NOT peek inside.
- **Slot 2 (Board 1) ALWAYS shows the character now in the product outfit, full body** — Cut 2 begins with her already in the product. Do NOT describe the changing motion. The hard cut from Cut 1 handles the implicit transition. For Board 1 specifically, embed the **twirl beat** inside Cut 2 (one brief natural twirl revealing the back of the outfit, then settles front-facing).
- **Slot 3 ALWAYS shows a FRONT_POSE** — Cut 3 reads the whole fit head to toe in an alive front-facing stance, on-camera dialogue (lip-sync).
- **Slots 4 and 6 ALWAYS show a HAND-FREE macro on a fabric / cut / detail beat** — each macro Cut includes ONE passive in-cut motion beat (breath, drape settle, light shift), NO hand contact with the fabric; Cut 6 lands on a DIFFERENT detail than Cut 4. Both are voiceover (character silent on camera).
- **Slot 5 ALWAYS shows a TURN** — Cut 5 shows the side or back of the garment mid-rotation, an over-the-shoulder glance, on-camera dialogue (lip-sync).
- **Slot 7 ALWAYS shows a styled pose in a different room** — Cut 7 lands the styled pose with on-camera dialogue (lip-sync) while the character holds it. Her mouth moves visibly with the Cut 7 portion of the monologue.
- **Slot 8 ALWAYS shows the FINAL_LOOK** — Cut 8 settles the loop-ready final wrap with on-camera dialogue (lip-sync).
- The story arc PRE_WEAR → WEARING → FRONT_POSE → TEXTURE_CLOSEUP → TURN → DETAIL → STYLE_POSE → FINAL_LOOK is the spine of Board 1; treat any deviation as an error in your reading of the board, not a creative choice.

**Critical reminder — board panels are SEQUENCE and TIMING reference only.** They confirm WHICH beat each slot represents (PRE_WEAR / WEARING / FRONT_POSE / TEXTURE_CLOSEUP / TURN / DETAIL / STYLE_POSE / FINAL_LOOK). They are NOT pose-by-pose frame templates. Your Cut description must invent the in-cut motion (breath, weight shift, fabric movement, twirl arc, expression evolution, hand mechanics) — these things are NOT on the static panel and must come from your text. Repeating the panel composition frame-for-frame in the Cut text gives Seedance two identical signals (image input + text caption) and produces stiff, lifeless output. The board says "this is the PRE_WEAR moment" — your text says HOW it unfolds in motion.

---

## Step 2 — POV Cadence and Style & Mood

Based on the board's per-slot POVs, set the Style & Mood line:

| Per-slot POVs | Style & Mood camera language |
|---|---|
| All eight slots SELFIE | `front-facing camera, intimate handheld feel` |
| All eight slots STATIC | `locked-off static camera, completely static, frozen frame` |
| POV varies between slots (e.g., SELFIE → STATIC → STATIC → STATIC-CLOSE → STATIC → STATIC-CLOSE → STATIC → STATIC) | `MIXED: starts [POV1] [language], then hard-cuts through the remaining beats — POV alternates per cut across the eight beats` |

The canonical Board 1 try-on cadence is `SELFIE → STATIC → STATIC → STATIC-CLOSE → STATIC → STATIC-CLOSE → STATIC → STATIC` (PRE_WEAR casual selfie with bag, then WEARING / FRONT_POSE / TURN / STYLE_POSE / FINAL_LOOK in static camera for full-body and styled framing, with the TEXTURE_CLOSEUP and DETAIL macros in static-close). Use the MIXED phrasing in Style & Mood for it.

For K > 1 boards, POV cadence depends on arc_role:
- `BOARD_2_TRY_ON_HOME_TOUR`: typically all STATIC with one or two SELFIE breaks (Cut 1 or a later pose beat) for variety. On-camera dialogue (lip-sync) on Cuts 1 / 2 / 3 / 5 / 7 / 8 regardless of POV; voiceover on the macro Cuts 4 and 6.
- `BOARD_3_TRY_ON_OUTDOOR`: typically all STATIC (outdoor locked framing); SELFIE OK on Cut 1 walking establishing. On-camera dialogue (lip-sync) on Cuts 1 / 2 / 3 / 5 / 7 / 8 — she lip-syncs while walking / standing / posing outdoors; voiceover on the macro Cuts 4 and 6 (wet-droplet fabric macros).
- `BOARD_4_TRY_ON_HOME_REFLECT`: Cut 1 default SELFIE (talking-head reflection opener) → Cuts 2 / 3 / 5 / 7 / 8 default STATIC (settled seated poses with on-camera dialogue, settled-glow register) → macro Cuts 4 and 6 STATIC-CLOSE voiceover (settled garment-detail macros)
- `BOARD_K_TRY_ON_LOOP`: match the K-type pattern for the cut's location

Choose POVs per cut action following Hand Allocation Rules.

---

## Step 2b — Baked camera moves (deliberate, opt-in)

Seedance renders the camera move you WRITE into the clip. OFF by default — add it only for a livelier edited-vlog energy or the "shot on a real phone" opener. A baked move is a single DELIBERATE, controlled dolly — never the banned uncontrolled `shake` / `drift` / `wobble` / `sway`, which stay forbidden everywhere.

- **Slow push-in / gentle ease-back.** Write it per cut: `the camera slowly PUSHES IN across the cut (a gentle dolly-in)` on a reveal / reaction beat, or `eases back and PULLS OUT` to open a wider beat. Echo it once in the quality tail (`a deliberate slow camera push-in then a gentle ease-back`).
- **On a LOCKED (static camera) cut, one deliberate slow push-in is the SOLE exception** to the freeze: the framing stays locked, the ONLY motion is that single intentional dolly (this is how a macro garment / texture beat gets its push-in) — still no shake / drift / wobble, and that cut's quality tail reads `locked framing with one deliberate slow push-in, otherwise static`.
- **Candid handheld ZOOM-IN opener (Cut 1, SELFIE).** For the real-phone opening, make Cut 1 a `candid HANDHELD iPhone ZOOM-IN toward the face — the frame pushes in fast and a little unsteady, a tiny overshoot-and-correct, like a real hand pinch-zooming, never a smooth professional dolly`; the first word / sound lands during the zoom. Echo in the tail (`an opening candid handheld iPhone zoom-in, then steady`).
- **At most ONE baked move per cut, never on every cut** — a move on every beat reads mechanical. Boards are stills and cannot encode motion; this is a clip-only instruction.

---

## Step 3 — Time-Slicing the Cuts

Distribute `clip_duration` across the 8 cuts. Default split for try-on — Cut 2 (the WEARING reveal with twirl on Board 1) and Cut 7 (the styled wrap pose) get slightly more time; the macro Cuts 4 (TEXTURE) and 6 (DETAIL) are briefer (single passive macro beat each); Cuts 1 / 3 / 5 / 8 are medium:

| `clip_duration` | C1 (PRE_WEAR) | C2 (WEARING) | C3 (FRONT) | C4 (TEXTURE) | C5 (TURN) | C6 (DETAIL) | C7 (STYLE) | C8 (FINAL) |
|---|---|---|---|---|---|---|---|---|
| 4s | 0.5s | 0.5s | 0.5s | 0.5s | 0.5s | 0.5s | 0.5s | 0.5s |
| 6s | 0.75s | 1s | 0.75s | 0.5s | 0.75s | 0.5s | 1s | 0.75s |
| 8s | 1s | 1.25s | 1s | 0.75s | 1s | 0.75s | 1.25s | 1s |
| 10s | 1.25s | 1.5s | 1.25s | 1s | 1.25s | 1s | 1.5s | 1.25s |
| 12s | 1.5s | 1.75s | 1.5s | 1.25s | 1.5s | 1.25s | 1.75s | 1.5s |
| 15s | 1.9s | 2.3s | 1.9s | 1.4s | 1.9s | 1.4s | 2.3s | 1.9s |

Adjust by ±0.25-0.5s if the action of a particular cut needs more or less time (the sum MUST equal `clip_duration` exactly). Each cut MUST remain ≥0.5s. At 4s every cut is the 0.5s floor (no room to weight); above 4s apply the weighting above.

For K > 1 boards (no twirl in Cut 2, all 8 cuts are K-specific beats of similar weight): use a more even split — divide `clip_duration` roughly equally across the 8 cuts, keeping the macro Cuts 4 and 6 slightly briefer. Exception: for `BOARD_4_TRY_ON_HOME_REFLECT` Cut 1 (talking-head opener) and Cut 8 (final wrap) may get slightly more time to land the reflection content; the macro Cuts 4 and 6 stay briefest.

Write the time spans into the Cut headers exactly (cumulative), e.g. for 15s: `Cut 1 (0-1.9s)`, `Cut 2 (1.9-4.2s)`, `Cut 3 (4.2-6.1s)`, `Cut 4 (6.1-7.5s)`, `Cut 5 (7.5-9.4s)`, `Cut 6 (9.4-10.8s)`, `Cut 7 (10.8-13.1s)`, `Cut 8 (13.1-15s)` — values per the table above.

---

## Step 4 — Action Language Per Cut

For each cut, write 4-10 sentences in the Dynamic Description describing the action. Rules:

### STATIC cut language
- Camera is **absolutely frozen and locked off — zero movement of any kind. No shake. No drift. No breathing wobble. No organic sway. No micro-movement. The frame is completely fixed and immovable. Only the subject and the product move within the locked frame.**
- **Locked camera does NOT mean a locked body.** Inside every STATIC cut the subject MUST visibly move within the frame — at minimum: a clear weight transfer, a mid-cut pose shift (e.g., hand-on-hip → arms-relaxed, or hand-in-hair → hand-on-thigh), a visible head turn or body rotation, plus continuous talking-head lip-sync motion (Cuts 1 / 2 / 3 / 5 / 7 / 8). A STATIC cut with the subject standing perfectly still = REWRITE. The locked frame is a frame for ALIVE posing, not a still photo.
- The Style & Mood / quality suffix MUST use locked-off STATIC phrasing for the static cut(s).
- **Forbidden words inside a STATIC cut's description:** `handheld`, `shake`, `drift`, `wobble`, `sway`, `slight movement`, `micro-shake`, `intimate handheld`, `natural movement`, `subtle movement`. These leak motion into the render. (These words describe CAMERA motion — subject motion words like `weight shift`, `pose swap`, `turn`, `step`, `roll`, `settle`, `lean` are encouraged and required.)

### SELFIE cut language
- **The phone is NEVER visible in frame.** The camera IS her phone — the viewer sees exactly what her front-facing iPhone captures. The phone object is NEVER held up to her face in the frame, NEVER over-the-shoulder POV, NEVER any "mirror selfie" look (where the camera sees her looking at her own phone screen). NO phone screen visible. NO third-person view of her holding a phone.
- Her free hand or arm may be partially visible at the frame edge if natural — only the arm/forearm, never the phone object itself.
- Natural handheld micro-shake from her grip is expected.
- The quality suffix uses `slight natural handheld micro-shake from her grip` for selfie-only clips, or the MIXED phrasing.

**Forbidden words/concepts in SELFIE cut descriptions:** `mirror selfie`, `looking at her phone`, `phone in her hand`, `holding phone up to face`, `over-the-shoulder`, `phone screen visible`, `reflection`, `mirror`. These leak phone-as-object or mirror-shot into the render — both are wrong for try-on.

### The 0.1-second hook law (EVERY clip, mandatory)

Cut 1 opens ALREADY MID-EVENT: frame one is mid-motion per the board's slot 1 — for Board 1 the character is already mid-action with the kraft bag (lifting it, turning with it, setting it down), NEVER a settled pose or a person waiting to start talking. And the voice starts IMMEDIATELY: the first spoken word (or the K=1 bracketed sound) lands within 0.0–0.4s of frame one — no silent lead-in, no breath-before-speaking, no settle-in beat. Write Cut 1's description so its FIRST clause is motion, and open the Audio line's first phrase at the very top of Cut 1. The only legal delay of the first word is an explicitly staged freeze-beat (hook staging only, ≤0.7s). In quiet voice-free mode the law transfers to sound: the first named SOUND lands ≤0.4s of frame one.

### H9 Entry Device (Cut 1 — OPTIONAL flavor, off by default)

Cut 1 may OPTIONALLY open on an operator-action hook — frame one reads as an accident of recording, not a directed shot. Optional flavor, never mandatory. At most ONE H9 device per clip; only when Cut 1's POV is SELFIE and `arc_role` is `BOARD_1_TRY_ON_CANONICAL` or `BOARD_4_TRY_ON_HOME_REFLECT` (the SELFIE talking-head opener). NEVER in a STATIC cut — the frozen-frame rules above admit no camera event. Never combined with a Set-Down / Pick-Up move (Step 4, Cut Markers) in the same cut.

Menu (pick ONE):
- **H9a Drop-Catch** — the frame is already tumbling as the clip opens, the world spins, then is caught and righted; the first word lands during the catch. Audio twin: fabric scrape on lens + a sharp audible breath.
- **H9d Walk-and-Slam** — fast handheld motion, she's walking in mid-stride (Board 1: the kraft bag handle at her side is native here), breathing audible, background streaking — the frame settles on her already talking. On Board 4 this stays settled-glow: she arrives and settles into the seat, no violent slam. Audio twin: footsteps + audible breathing into room tone.
- **H9g Light Switch** — near-black, only her voice; a lamp clicks on and the scene appears already mid-moment — sound leads picture by half a second. Audio twin: one lamp click, a single punched event.
- **H9e Zoom-Out Reveal** (Board 1 only) — extreme close digital zoom on the kraft paper texture of the bag; a quick zoom-out reveals the full Cut 1 staging; the first line refers to what was just revealed. Audio twin: muffled room tone opening to full room tone on the zoom-out.

Laws:
- **FRAME-PHYSICS language only** — the frame tumbles / swings / settles / clears. The SELFIE forbidden words above stay in FULL force inside an H9 description: never "drops her phone", never "holding phone", no phone-as-object words; `reflection` / `mirror` stay banned — no H9 via mirrors.
- **The event resolves INTO the board's slot-1 composition within ~1.5s** — the board panel is still the moment the camera arrives at. Cut 1's staging laws stand unchanged: pre-wear outfit + kraft bag as contextual prop on Board 1 (no opening, no peeking), seated talk-opener on Board 4.
- **The first spoken word still lands ≤0.4s in / during the event** — Step 6's first-word and opener rules stand; this device EXTENDS them. K=1 bracketed non-verbal sounds may lead as usual (they don't count as first words).
- **Every H9 device carries its audio twin written into the Audio line's ambient clause** (same precedent as the room-tone / fabric-rustle ambience) — a silent camera event renders as a glitch.

### Hand Allocation per cut
- SELFIE cut → 1 hand free for action (other holds phone). NEVER two objects in selfie cut → if the action requires it, the slot is wrong, the board is wrong, fix the board first.
- STATIC cut → 2 hands free. Suitable for natural posing, two-handed adjustments, full-body framing.
- **Cuts 4 and 6 specifically (the macro cuts) — NO hand contact with the product fabric** regardless of POV. Hands stay at sides / off-frame / clearly NOT touching the garment.

**THE HAND-COUNT LAW (hard cap — one hand or two, NEVER three).** The character has exactly TWO hands, and every hand written into a cut belongs to her. At any simultaneous moment of a cut, the hand roles described (and expected in frame) total ≤ 2 — an action load that needs a third hand ("holds the bag while adjusting the collar while waving") writes a phantom third arm into the render; a third hand must be impossible to read out of the prompt. The law composes with the bullets above: SELFIE's one-free-hand cap, STATIC's two-free-hands, and the macro cuts' (4 and 6) hand-free macro (the ultimate park) stand unchanged. The garment WORN on the body is NOT a hand role.

- **Name the role of EACH hand.** One-hand actions name the acting hand AND park the other explicitly (in SELFIE the parked hand IS the phone grip, off-frame; in STATIC: resting at her side, on her hip, flat on the counter). Two-hand actions are LEGAL when the action naturally needs both (a two-handed accessory adjustment per Weight & Grip, holding the kraft bag by both handles) — then name both roles in ONE sentence ("left hand steadies the bag on the console, right hand rests on the handle") and give the hands NO other simultaneous job.
- **Third-hand-prevention staging:** prefer one hand actively on the prop with the other parked; when a prop would otherwise need a stabilizing hand, rest it on a surface instead (the Cut 1 kraft bag set upright on a premium surface is the native example). A prop floating unheld beside busy hands also spawns the third hand — it is held or it is resting, explicitly. Multi-step actions sequence across the existing hard cuts, never piled into one beat.
- **Count before output:** before finalizing, count the hand roles written into each cut — total simultaneous ≤ 2.

### Bag Presence per Cut (mandatory for try-on)

The try-on centers on a kraft paper shopping bag. Each Cut of Board 1 has a fixed bag state:

- **Cut 1 (PRE_WEAR)**: kraft paper shopping bag visible — plain brown craft paper, no logo, optional handles tinted to match the product's primary color. Character is in pre-wear outfit. Product is NOT visible. Bag is held by one handle at her side OR set upright on a premium surface beside her. **Character does NOT open the bag, does NOT lift the product out of the bag, does NOT peek inside, does NOT vlog the bag.** The bag is a contextual prop — that's it. Hard cut to Cut 2.
- **Cut 2 (WEARING)**: kraft bag GONE from frame entirely. Character now in the product outfit, full body, locked static camera. **Do NOT describe the act of changing.** The hard cut from Cut 1 handles the implicit transition. Cut 2 begins with her already in the product. For Board 1 (twirl applies), embed ONE brief natural twirl revealing the back of the outfit, then she settles front-facing.
- **Cut 3 (FRONT_POSE)**: kraft bag GONE. Character in product outfit, front-facing, reading the whole fit head to toe in an alive stance.
- **Cut 4 (TEXTURE_CLOSEUP)**: kraft bag GONE. Character in product outfit. HAND-FREE macro on fabric / cut / detail — see Step 4b below for the strict no-hand-contact rules.
- **Cut 5 (TURN)**: kraft bag GONE. Character in product outfit, mid-turn showing the side or back of the garment.
- **Cut 6 (DETAIL)**: kraft bag GONE. Character in product outfit. Second HAND-FREE macro on a DIFFERENT detail than Cut 4 — see Step 4b below.
- **Cut 7 (STYLE_POSE)**: kraft bag GONE. Character in product outfit, in a DIFFERENT room of the same home, settled into a styled pose.
- **Cut 8 (FINAL_LOOK)**: kraft bag GONE. Character in product outfit, settled alive final wrap (STYLE_POSE room or a natural continuation), loop-ready.

Once the bag has disappeared after Cut 1, NEVER re-introduce it in subsequent Cuts. No re-handling, no re-setting on a surface, no second bag. The bag ceases to exist.

### Per-K Cut behavior (K > 1)

For K > 1 the kraft bag NEVER appears, the pre-wear outfit NEVER returns, character / hairstyle / product outfit stay locked. Each K has its own per-cut behavior:

- **`BOARD_2_TRY_ON_HOME_TOUR` (K=2):** Cut 1 bridges from Board 1's Slots 7-8 location vibe — same or adjacent room, lighting continuation, character in product outfit, body settles / transitions into a new pose (small body shift, glance around the room). Cuts 2-8 are pose variations in DIFFERENT home rooms (kitchen / hallway / bedroom / balcony / living room — any room not yet shown on Board 1). The macro Cuts 4 and 6 are hand-free macros on garment details (hem / sleeve / collar / hardware / seam — a different detail each) per Step 4b. No twirl beat. On-camera dialogue (lip-sync) on Cuts 1 / 2 / 3 / 5 / 7 / 8 regardless of POV — mid-thought continuation, no greetings. Macro Cuts 4 and 6 voiceover layered (silent on camera).
- **`BOARD_3_TRY_ON_OUTDOOR` (K=3):** All 8 cuts outdoor, tier-matched. Cut 1 = dry outdoor establishing (character walking or standing, full body, no rain). Cut 2 = light rain begins — character in light drizzle, water droplets visible on fabric (shoulders / arms / hem), pavement may show wet sheen. Cut 3 = front-facing outdoor fit read in the drizzle. Cut 4 = HAND-FREE macro on wet droplets beading on the fabric surface — no hand contact (per Step 4b). Cut 5 = TURN showing the side / back mid-walk. Cut 6 = second HAND-FREE wet macro on a DIFFERENT detail than Cut 4. Cut 7 = post-rain or sheltered styled pose (under awning / café terrace / doorway), droplets may persist on fabric. Cut 8 = settled outdoor final wrap. Hair stays dry across all 8 cuts. No twirl beat. Register-matched energy in outdoor body language (walking-confident, standing wet-confident — natural engaged by default; hyped body language only on the explicit energy signals). On-camera dialogue (lip-sync) on Cuts 1 / 2 / 3 / 5 / 7 / 8 — she lip-syncs while walking, standing in the drizzle, turning, landing the wrap pose. Macro Cuts 4 and 6 voiceover layered (fabric macro framing).
- **`BOARD_4_TRY_ON_HOME_REFLECT` (K=4):** All 8 cuts back inside the home (1-2 tier-matched rooms). Character settled (seated / lounging — sofa / armchair / window-seat / bed / kitchen island). Cut 1 = SEATED_TALK_OPENER. **Cut 1 default POV is SELFIE → on-camera dialogue lip-syncing the Cut 1 audio segment** (talking-head reflection vibe). Cuts 2 / 3 / 5 / 7 = settled poses at different angles / adjacent rooms, STATIC, on-camera dialogue (lip-sync) in settled-glow register. The macro Cuts 4 and 6 = hand-free macros on hem / sleeve / collar at rest (a different detail each), voiceover layered. Cut 8 = final settled wrap, STATIC, on-camera dialogue (lip-sync) from the final settled pose. Expression stays in **settled-glow register** (sustained warm grin, contented victorious energy, body calm-settled — NOT explosive-scream) whatever the clip's register.
- **`BOARD_K_TRY_ON_LOOP` (K≥5):** 8 pose variations alternating previously-shown / new tier-matched home rooms or outdoor spots. Cut behavior matches the slot's location type (indoor → K=2/K=4 cut rules; outdoor → K=3 cut rules; settled reflection → K=4 rules); the macro Cuts 4 and 6 stay hand-free voiceover.

### Outfit Continuity per Cut (mandatory)

The try-on hinges on the pre-wear → product outfit transition.

- **Cut 1 (Board 1 only)**: character is in the **pre-wear outfit** — explicitly describe: `She wears a [pre-wear description — basic tee + lounge pants / oversized hoodie + cotton shorts / simple robe / plain knit + sweatpants — boring/neutral home base].`
- **Cut 2 onward (Board 1) + every Cut of Boards 2..N**: character is in the **product outfit**. The pre-wear outfit is GONE forever after Cut 1 of Board 1 — never re-introduce it.
- **Hairstyle stays consistent across all cuts and boards.**

**No on-screen costume change:** the transition between Cut 1 (pre-wear) and Cut 2 (product) is implicit via the hard cut. NEVER describe:
- The character opening the bag
- The character changing clothes on screen
- The character pulling the product out of the bag
- The character putting the product on
- The character lifting fabric over her head / pulling a garment down / zipping anything

The transition is invisible to the viewer — she's in pre-wear, hard cut, she's in product. That's the magic of the try-on cut.

### Twirl Beat (Board 1 / Cut 2 only)

For Board 1 (`arc_role == "BOARD_1_TRY_ON_CANONICAL"`), Cut 2 MUST embed ONE brief natural twirl beat revealing the back of the outfit. Pattern:

`In roughly the middle of the cut, the character takes a slow exhale, weight shifts onto the right foot, and she rolls her body in one smooth turn revealing the back of the outfit (hold for half a beat as the back is visible), then turns back to camera and settles, hands relaxed at her sides. The [garment fabric type — silk / linen / denim / knit / leather] catches a slight breath of motion as she turns — natural drape, not exaggerated. The rotation is ONE single turn, not multiple, not continuous, not a slow spin — a quick natural reveal motion.`

For ALL K > 1 boards (`BOARD_2_TRY_ON_HOME_TOUR` / `BOARD_3_TRY_ON_OUTDOOR` / `BOARD_4_TRY_ON_HOME_REFLECT` / `BOARD_K_TRY_ON_LOOP`): NO twirl. Cut 2 is its K-specific beat per the Per-K Cut behavior block above — no rotation, no reveal twirl. Describe the cut's beat in motion, micro-expressions, natural settle.

### Garment Consistency Lock (mandatory)

When `product_media_id` is provided, Garment Consistency Lock applies across every Cut in which the product is worn (Cuts 2-8 of Board 1; all Cuts of Boards 2..N):

`The garment keeps identical silhouette, primary color, print, and recognizable design details (collar style, hem, sleeve, neckline, hardware, stitching) across every Cut in which it appears. The character may turn or pose freely — the garment rotates naturally with her body. The product is rendered at realistic real-world proportions on the character's body — natural drape per the fabric weight, natural fit, not exaggerated.`

### Weight & Grip Logic (accessories only)

For apparel (the typical try-on case), weight logic is not relevant — the character simply wears the garment. For accessory products, classify by weight before describing the holding/placing action:

| Class | Examples | Hand allocation | Facial expression |
|---|---|---|---|
| Heavy | Large structured handbag, boots being put on | TWO hands required, body leans slightly | Slight composed effort combined with the register-matched try-on reaction (genuine delight by default; hyped only on the explicit energy signals) |
| Bulky but light | Oversized hat, large scarf | TWO hands for adjusting | NO strain — relaxed face, easy grip |
| Light | Sunglasses, small clutch, jewelry, watch | ONE hand or natural placement | Neutral / register-matched lively depending on the slot's expression beat, no strain |
| Tiny | Earring, ring | Pinched (thumb + index), close to lens | Focused / curious, no strain |

**Forbidden:** describing one-handed lifting of heavy accessories, or two-handed strain on light accessories. **Paired accessories (earrings sold as pair):** never both halves balanced on a single palm — one in each hand, or one displayed plus one set down, or side-by-side on a flat surface.

---

## Step 4b — Hand-Free Macro Logic (Cuts 4 and 6 specifically)

The macro Cuts 4 and 6 (TEXTURE_CLOSEUP + DETAIL for K=1; FABRIC_WATER_MACRO ×2 for K=3; SETTLED_GARMENT_DETAIL ×2 for K=4; garment-detail macros for K=2) show the product's fabric / cut / detail through framing and natural body movement, NOT through hand interaction. Cut 6 always lands on a DIFFERENT detail than Cut 4. **NO touching, NO skimming, NO pulling, NO brushing, NO pinching** — hands stay at sides / off-frame / clearly NOT in contact with the close-up garment area. This rule applies on ALL boards' Cuts 4 and 6.

The texture / cut / detail reads through ONE passive in-cut motion beat per macro Cut, picked by garment type (pick a different one for Cut 6 than Cut 4):

| Garment type | Cuts 4 / 6 hand-free in-cut motion (pick ONE per macro) |
|---|---|
| Top / shirt / blouse / tee | Light catches the chest area as she takes a slow breath, the surface plays in soft directional light / shoulder seam shifts subtly as the body settles / fabric falls naturally over the chest with slight body sway |
| Dress | Skirt drapes naturally as the body settles / waist seam shifts visibly as she breathes / bodice catches light across the panel as the body settles |
| Skirt | Skirt drape settles in the locked frame / side seam shifts subtly / fabric reads pleat structure as light shifts |
| Pants / shorts | Fabric falls along the leg with natural drape as she shifts weight / cuff sits naturally / seam line visible as the body settles |
| Outerwear / jacket / coat | Lapel sits open with collar settled / shoulder seam reads as the body settles / fabric texture catches the light |
| Knitwear / sweater | Knit texture catches light as the body breathes / shoulder seam visible / weave reads clearly in macro framing |
| Denim | Wash gradient visible as the leg shifts / seam line reads / fabric grain catches light |
| Leather | Grain catches light / sheen reads naturally as the body settles / surface plays in directional light |
| Accessories | Catches light / sits naturally as the body settles / detail reads in macro framing |

### K=3 wet-droplet macro motion (Cuts 4 and 6 of `BOARD_3_TRY_ON_OUTDOOR`)

When Cuts 4 and 6 are the FABRIC_WATER_MACROs on K=3, pick ONE wet-droplet in-cut motion per macro (a different one for Cut 6 than Cut 4):

| Garment type | K=3 Cuts 4 / 6 wet macro in-cut motion |
|---|---|
| Silk / satin | A single droplet rolls across the surface and slides off the edge of the frame |
| Linen / cotton | Droplets bead and sit on the surface, weave clearly readable as the body breathes |
| Leather | Droplets bead high like glass, leather grain catches directional light as the body settles |
| Denim | Droplets bead on the heavier denim surface, wash gradient unchanged, fabric grain clear |
| Knitwear / sweater | Droplets nestle into the weave texture without soaking, knit pattern reads clearly |
| Synthetic / waterproof | Water beads as if on a windshield, full surface sheen catches light as the body shifts |
| Outerwear / jacket / coat | Droplets bead on the outer shell, one droplet slides along a seam |

The macro stays HAND-FREE — character's hands at sides / off-frame.

### K=4 settled garment-detail macro motion (Cuts 4 and 6 of `BOARD_4_TRY_ON_HOME_REFLECT`)

When Cuts 4 and 6 are the SETTLED_GARMENT_DETAILs on K=4, pick ONE at-rest in-cut motion per macro (a different one for Cut 6 than Cut 4):

| Garment area | K=4 Cuts 4 / 6 settled macro in-cut motion |
|---|---|
| Hem (skirt / dress / coat) | Hem rests in lap or drapes over the seat edge, fabric falls naturally as the body breathes |
| Sleeve cuff | Cuff sits at wrist or over armrest, light catches the seam as the body settles |
| Collar / lapel | Collar sits open at the neckline, fabric texture reads clearly in close framing |
| Knit weave | Knit catches indoor lamp-light or window-light as the body breathes |
| Fabric drape on lap | Fabric folds in lap, drape settles naturally as the body breathes |

Hands stay off the garment — character's hands rest on armrest / lap-edge / off-frame.

General rules:
- ONE hand-free in-cut motion beat per Cut. Never multiple, never back-and-forth.
- The character's hands stay at her sides / off-frame / behind the back / clearly NOT in contact with the close-up garment area.
- The garment stays in place — only natural drape / light / breath movement.
- **NO "operator hand"** — no hand from outside the character's body enters the macro frame.
- The character does NOT touch her own clothing in the close-up framing — even gentle brushing reads as "someone touching her clothes during filming" in the rendered video.

---

## Step 5 — Cinematic Specificity (mandatory per cut)

Each cut must include all three of:

1. **5+ concrete micro-beats** from the register-matched menu below (rotate — never repeat the same combination across the 8 cuts). **Default emotional register is NATURAL — conversational, lively, genuinely engaged; pick predominantly from the natural menu.** Switch to predominantly-hyped picks ONLY when `user_request` carries an explicit energy signal (`hyped` / `hype` / `energetic` / `explosive` / `high-energy` / `viral energy` / `insane energy`). Switch to predominantly-calm picks ONLY when `user_request` explicitly signals one of: `goth`, `vampire`, `cinematic noir`, `cold`, `passive`, `deadpan`, `clinical`, `refined`, `luxury-passive`, `minimal`, `somber`, `serious`, `dark`, `shadowy`, `posh-restrained`, `runway-cool`, `quiet`, `GRWM`, `routine`, `process-led` tone or aesthetic. Keep the energy language concrete in every register — Seedance under-renders energy; a flat-neutral prompt renders a wooden AI presenter.

   **Natural menu (default — pick from here first):** raised brows with a genuine grin, small bright laugh, lean-in toward the lens, head tilt appraising the fit, surprised blink, satisfied slow nod, half-laugh through the nose, breaking grin she doesn't fight, delighted glance down at the outfit then back up with a warmer smile, one hand smooths the hem at the waist (never in the macro cuts 4 / 6), small proud shoulder settle, honest jaw-drop that relaxes into a smile, weight rock back with a pleased exhale, quiet appreciative head shake.

   **Hyped menu (opt-in — pick from here first ONLY on the explicit energy signals above):** WILD open-mouth grin (jaw dropped wide, eyes blown wide, full-body delight), mouth blown open in scream-laugh of excitement, victorious open grin, dramatic head jerk back with explosive joy, cheek puff out then deflate, eyebrows shoot skyward, knuckles soft-press into hip with grin-tighten, mock-confused squint then break-into-laugh, slow head shake with massive grin, full-body satisfaction shudder, tongue-press inside cheek, eye-roll then explosive grin back to lens, head thrown back with burst of laughter, fingers fan out beside the face in "I can't believe" gesture.

   **Calm menu (override — pick from here only when `user_request` signals calm tone):** weight shift, hair touch, glance break, head tilt, eyebrow flash, hand gesture (NOT on garment in the macro cuts 4 / 6), posture shift, lip movement, shoulder roll, breath (inhale / exhale / sigh / sharp inhale), jaw set, neck tendon definition, foot pivot, brow furrow, chin tuck, lean forward / back, micro-grin, half-blink, slight off-center handheld tilt.

   **Quiet process-led mode** (on a `quiet` / `GRWM` / `routine` / `process-led` signal in `user_request`): the calm override applies in full, plus — a sparse `monologue_segment` is LEGAL (≤ 20 words per 15s clip; the word-density floor is waived — distribute it thin, never pad; a voice-free cut is legal under this calm-tone override); SOUND carries the clip instead: denser precisely named SFX in the Audio line's ambient clause (fabric slide, hanger clicks, zips, taps — close and intimate). An early Set-Down is the preferred transition (the Set-Down / Pick-Up rules apply unchanged — Board 1 Cut 1→2 stays a hard cut, always); no presentation gestures — the garment enters as one honest routine step; keep ONE tiny human beat so it breathes. The 0.1s hook law holds: when Cut 1 opens voice-free, the first named SOUND lands ≤0.4s of frame one instead of the first word. In this mode the playful-improv mandate collapses to the ONE tiny human beat (no separate goofy beats), and on the closing clip (K == N) the default Pick-Up wins over the early Set-Down preference (max one device stands).

2. **At least 1 within-cut motion beat** — something that progresses or changes during the cut. The cut is not a still — describe what evolves inside it. Examples for try-on: `weight shifts forward as she psychs herself up off-camera`, `she takes a small breath as the new outfit settles on her shoulders`, `the silk catches a slight breath of motion as she completes the twirl`, `she settles into the chair, exhales softly, and gives one small confident nod`, `a quick genuine grin breaks across her face after the controlled exhale`.

3. **Expression evolution across the 8 cuts** — never the same expression twice (NATURAL default: bright genuine anticipation on the bag (C1) → genuine wide-eyed delight on the new look (C2) → warm confident read of the full fit (C3) → engaged admiration on the texture detail (C4) → pleased curiosity through the turn (C5) → quiet admiration on the second detail (C6) → proud victory on the styled pose (C7) → satisfied confident settle on the final look (C8). HYPED — only on the explicit energy signals: hyped anticipation on the bag → peak hyped reaction on the new look → hyped fit-check → hyped admiration on the detail → hyped turn → hyped second-detail beat → victory celebration on the styled pose → sustained hyped settle). Identical expression across cuts is forbidden.

### Per-K micro-beat hints (mandatory for the relevant K)

In addition to the menu above, weave in K-specific environmental / contextual beats:

- **K=2 home tour:** glance around the new room mid-cut, eyes drift across furniture / window light, hand grazes a doorframe or armrest as she moves through the room, small confident step into the new space.
- **K=3 outdoor:** wind catches the fabric drape mid-cut, rain droplets visible on shoulder or hem, foot-step on wet ground audible in motion, character's collar adjusts in breeze, glance up at the sky as drops fall, slight head-shake (NOT shaking water out of hair — just confidence).
- **K=4 home reflect:** settles deeper into the seat, eyes drift around the room (catching herself in the corner of her eye), hand rests on fabric or armrest (allowed on K=4 outside the macro cuts 4 / 6), small grin breaks while reflecting, slow exhale of contented satisfaction, sip of tea / glance at her own outfit in lap.

These K-specific beats COUNT toward the 5+ micro-beats requirement — they don't add on top.

**Micro-beat writing discipline:** name the body part and the object (`tucks a strand behind her ear with one finger`, never `fidgets`); ONE movement at a time — simultaneous movements read as glitching; place beats BETWEEN spoken phrases, never on a key word (they smear lip-sync; they double as resync anchors); at least 3 of the 5+ are distinct named-body-part movements; every audible beat (tap, rustle) goes into the Audio line or renders mute; movements carry the emotion — never write the emotion word.

**Peak reactions are BODY events — 1-2 per clip max, never three.** The default arc's peak is Cut 2 (the new-look reveal); an optional second may land on Cut 7's victory beat. Every peak is product-motivated (reveal / detail / styled landing), never a random scream. In the NATURAL register the peak stays at HUMAN scale — a real jaw-drop, a breaking grin, a delighted laugh — genuine, never theatrical, never staged screaming; screaming peaks belong to the HYPED register only. Pair the vocal moment with exactly ONE body event (sharp gasp with hand to chest, jaw-drop that STAYS, snap lean-back) — it replaces one micro-beat slot, never stacks. Quirk / goofy beat and peak NEVER share a beat — separate timestamps, or both render as mush.

**Performance tendency — at least 1 unguarded micro-beat per clip.** Real UGC creators break character, recover, and let micro-mistakes through. Include at least one recovered eye-flick / mid-thought stumble / post-laugh settle / quick self-correction / re-found composure / "wait what was I saying" beat. Wooden, posed-throughout performances read as AI. Skip this tendency only when `user_request` explicitly specifies a sustained deadpan / clinical / cold tone that should hold across the whole clip.

**Sound intrusion (optional — max ONE per clip, use when the location makes an off-frame sound natural).** The world interrupts. Three parts, always: (1) the sound goes into the Audio line's ambient clause (same precedent as the room-tone / fabric-rustle ambience), named precisely ("a dog barking close, two bursts", "a kettle starting to whistle off-frame" — never "a noise"); the sound MUST belong to the location (home rooms on K=1/2/4, street sounds on K=3 outdoor). (2) A physical reaction beat: eyes flick off-frame toward the sound, a half-turn of the head, one beat of held stillness — a glance, never a hand action (the hand-count law stands). NO new spoken words — the monologue stays verbatim; the voice pauses briefly at a phrase boundary or she keeps lip-syncing over it with a slight frown, never acknowledges the sound in words. (3) The return: back to the lens within ~1.5s. The intrusion SHARES the unguarded-beat slot above — it can BE the clip's unguarded beat, never stacks on top; if the clip already has a strong unguarded beat, skip it. Never during a peak (Cut 2's reveal / twirl), the Cut 7 wrap landing, or the macro cuts (4 / 6) hand-free macros — park it in a talking-head body beat.

**Playful improv tendency — at least ONE small goofy beat in EACH of at least THREE different on-camera cuts (mandatory; pick from Cuts 1 / 2 / 3 / 5 / 7 / 8 — e.g. 1, 3, 7).** Real try-on creators ham it up across the whole clip — they pull mock faces, do little physical bits, break the "selling" frame for a half-second every time the camera catches them. Lean into natural creator goofiness — pick a DIFFERENT beat for each chosen cut so the clip doesn't repeat itself. Menu (not exhaustive, pick whatever fits the moment): tongue-out flash, "blep" face, mock-zen closed-eyes, eyebrow waggle, exaggerated mock-thinking face with finger on chin (only outside the macro cuts 4 / 6 — never on a macro), double thumbs-up with cartoon grin, mid-gesture cartoon shrug, mock-disappointment slow head shake, sudden mock-serious lip-purse that breaks into a grin, finger-guns at lens, mock-fan-yourself wave, hand-over-heart mock-faint, mock "chef's kiss" gesture, mock-bow with the styled pose. Natural-improv, never theatrical — a half-second beat woven into the cut's flow, not a posed performance.

- **The macro cuts (4 and 6) are EXEMPT** from the playful improv mandate — the hand-free fabric macro centers on garment detail, face is often partially or entirely out of frame, and hands stay off the garment. Skip the goofy beat in the macro cuts.
- **K=4 (HOME_REFLECT) register modifier:** the three goofy beats stay in **settled-glow** register — softer picks like mock-thinking finger-on-chin, eyebrow waggle, soft mock-bow head-tilt, contained chef's-kiss, small mock-serious lip-purse breaking into a grin. AVOID explosive picks (no cartoon double-thumbs-up, no big mock-fainting moves) — they break the settled posture.
- **Skip ALL goofy beats** only when `user_request` specifies a sustained `refined` / `clinical` / `cold` / `luxury-passive` / `goth` / `vampire` / `cinematic noir` / `somber` / `serious` / `posh-restrained` / `runway-cool` / `quiet` / `GRWM` / `routine` / `process-led` tone where playfulness would break register.
- **Detected signature quirk (from `user_request` only):** a named repeatable quirk (trust-tap, glasses push, whisper-echo) REPLACES one goofy-beat slot — never a 4th beat, never in a macro cut (4 or 6), never on a peak's beat. Stage it LARGE (small-written quirks render ~2 of 9): its own timestamped window (up to 2.0s where the cut allows), nothing else in it, mechanics 30% bigger, the body part filling its frame zone, sound punched in the Audio line as a SINGLE event (`one sharp loud fingernail tap`, never `taps` — repeated taps trigger motion loops). Same quirk, same trigger, verbatim on every board. A detected quirk is never silently dropped: it survives the skip-ALL-goofy-beats tone gate (staged in the requested register); all other quirk rules stand.

**Forbidden in any Cut description:** sentences that only re-state what the static board already shows. Every sentence must add something the board cannot — motion, breath, fabric movement, expression beat, kinetic detail.

Anti-patterns (NEVER write these):
- "smiles at the camera"
- "looks at the camera"
- "sits in front of the camera wearing the outfit"
- "poses for the camera"
- "holds the product and talks"
- Identical expression across all 8 cuts

Instead: weave specific micro-behaviors into each cut's description.

### Cut Markers (mandatory verbatim)

Between every consecutive pair of cuts (Cut 1→2, 2→3, 3→4, 4→5, 5→6, 6→7, 7→8): `Hard cut to.` at the end of each cut's description sentence — SEVEN markers total.
No marker after Cut 8.

These are scene-edit instructions Seedance reads literally. Without them, cuts collapse into smooth motion. (ONE optional exception exists — the Set-Down / Pick-Up device below.)

The Cut 1 → Cut 2 hard cut implicitly handles the costume change (pre-wear → product). Never describe the changing motion in either Cut 1 or Cut 2.

### Set-Down / Pick-Up (replaces exactly ONE `Hard cut to.`)

Default policy: on the CLOSING board (`K == N`), when the Cut 6 → Cut 7 boundary is legal (Cut 7 on-camera SELFIE, no state jump, Cut 7 span ≥ 3s), USE the Pick-Up instead of that boundary's `Hard cut to.` — her off-camera Cut 6 voice carries through the move and lip-sync engages the moment her face lands in the handheld frame; the styled wrap lands at arm's length, face close and slightly wide-angled. On other boards the device stays off unless `user_request` signals a one-take / single-take / honest-take feel — then use one device per clip wherever a legal boundary exists. If `user_request` asks for classic hard cuts, hard cuts everywhere. When active, EXACTLY ONE `Hard cut to.` per clip is replaced — every other boundary keeps `Hard cut to.` verbatim. Max one Set-Down OR Pick-Up per clip, never both, and never combined with an H9 entry device in the same cut.

- **Set-Down** (only between an adjacent SELFIE cut → STATIC cut): still mid-sentence, she lowers the phone below the frame line — the frame swings down, tilts, and settles leaning at a slight low angle, slightly crooked. A perfectly level set-down reads fake. She steps back into full view, hands now free, and keeps talking without a pause. After the set-down the frame is static — the STATIC cut language applies, no handheld words. The phone object itself is NEVER visible in frame.
- **Pick-Up** (only between an adjacent STATIC cut → SELFIE cut; strongest right before the closer cut): she walks toward the camera and **reaches past the lens** — never "grabs the phone" — the frame lifts, shakes for a beat, and becomes handheld again, her face now close and slightly wide-angled. Handheld micro-shake returns per the SELFIE rules.

Laws:
- **HARD LAW — never across an implicit state jump.** On Board 1, Cut 1 → Cut 2 is FORBIDDEN for this device: that hard cut is what hides the outfit change (PRE_WEAR → WEARING) — a continuous transition would render the change on camera. That boundary keeps `Hard cut to.` verbatim, always.
- The natural legal spot in this flow is a macro Cut → lip-sync Cut boundary (Cut 6 → Cut 7, or Cut 4 → Cut 5): the macro Cut is voiceover (character silent on camera), so a Pick-Up carries her off-camera voice through the move and lip-sync engages the moment her face lands in the handheld frame — the next cut's on-camera dialogue. (Board 4's Cut 1 SELFIE → Cut 2 STATIC is a legal Set-Down boundary — no state jump there.)
- **The voice runs THROUGH the move** — audio continuity is what sells the one-take. Note it in the Audio line at the boundary (voice unbroken through the transition).
- The transition beat costs ~1.5s — it must fit inside the cut's time span (Step 3) without starving the monologue distribution across the 8 cuts.

---

## Step 6 — Audio / Monologue (Try-On Audio Model)

Use the provided `monologue_segment` verbatim. Distribute it across the 8 cuts at natural phrase boundaries — roughly proportional to cut duration. When phrase boundaries allow, bias the wordiest chunks toward the voiceover macro cuts (4 and 6) — they're voiceover, so there are no lips to slop.

### Performed delivery markup (words stay verbatim)

The words are untouchable; the DELIVERY is yours. Sanctioned stylings (alongside the opener rewrite and dedup): a **stretched vowel** at a peak (`so` → `soooo`) — max 2 per clip, each eats ~1.5s; a **CAPS spike** on max 1-2 words per cut segment (caps everywhere = shouting robot); **ONE broken sentence at the peak** — a single em-dash break; **max ONE whisper-to-spike switch per clip**, scripted in the Audio line (near-whisper lead-in, spike on the reveal). **Breath events are timed actions** — written at their moment in the cut text AND in the Audio line as sound (`a sharp audible gasp, hand flying to her chest`); K=1 may also use the bracketed pool below; K>1 keeps the NO-bracketed-sounds rule (described actions + Audio ambience only). On K == N, the loop rule's no-text-edit applies on top of this markup — no styling may alter the final phrase used for the mid-phrase timing cut.

**Try-on audio model — alive talking-head register:**

The character SPEAKS ON CAMERA across Cuts 1, 2, 3, 5, 7, and 8 — mouth visibly moves, lip-syncing the corresponding monologue portion in each cut. Cuts 4 and 6 (the hand-free fabric/detail macros) are the voiceover cuts, because the framing centers on garment detail rather than the face. This split keeps the talking-head energy alive throughout the try-on while preserving the macro-detail integrity of the macro cuts.

- **Cut 1 (PRE_WEAR)** — on-camera dialogue regardless of POV. Character speaks visibly to lens, mouth moves, lip-syncing the Cut 1 portion. Works for both SELFIE (intimate handheld) and STATIC (locked-off, three-quarter framing — she still lip-syncs to the locked camera across the room).
  - K=1 (BOARD_1_TRY_ON_CANONICAL): default POV SELFIE → on-camera dialogue introducing the moment.
  - K=2 (BOARD_2_TRY_ON_HOME_TOUR) and K=3 (BOARD_3_TRY_ON_OUTDOOR): typically STATIC POV → still on-camera dialogue mid-thought continuation, no greetings.
  - K=4 (BOARD_4_TRY_ON_HOME_REFLECT): default POV SELFIE → on-camera dialogue, talking-head reflection in **settled-glow register** ("now that I've actually worn this" / "okay so I've been in this all day").
  - K≥5 (BOARD_K_TRY_ON_LOOP): on-camera dialogue mid-thought; POV per location type.
- **Cut 2 (WEARING)** — on-camera dialogue. Default POV STATIC. Character is now in the product outfit, full-body / three-quarter framing, locked-off camera across the room — she lip-syncs the Cut 2 portion while moving through her pose (and the Board 1 twirl beat lands inside the cut). For Board 1, the lip-sync momentarily pauses during the twirl reveal motion and resumes the moment she settles front-facing — the mouth does NOT form words while her back is to camera mid-rotation.
- **Cut 3 (FRONT_POSE)** — on-camera dialogue. Character front-facing, reading the whole fit head to toe, lip-syncing the Cut 3 portion at a different distance/POV from Cut 2.
- **Cut 4 (TEXTURE_CLOSEUP)** — **voiceover layered.** Character is silent on camera, mouth does NOT form words / does NOT lip-sync. Her voice plays as off-camera narration speaking the Cut 4 portion. The framing is a hand-free macro on fabric / cut / detail — face is partially or entirely out of frame, so on-camera dialogue would not register.
- **Cut 5 (TURN)** — on-camera dialogue. Character mid-turn showing the side / back of the garment, an over-the-shoulder glance, lip-syncing the Cut 5 portion (the lip-sync momentarily pauses only if her back is fully to camera mid-rotation, resumes as she faces back).
- **Cut 6 (DETAIL)** — **voiceover layered.** Character silent on camera, mouth not forming words. Her voice narrates the Cut 6 portion off-camera. Second hand-free macro on a DIFFERENT detail than Cut 4.
- **Cut 7 (STYLE_POSE)** — on-camera dialogue. POV per the cut's framing — STATIC (default for full-body styled wrap) or SELFIE (when the styled pose is naturally one-handed, e.g. walking with phone, leaning with one hand free). Character lip-syncs the Cut 7 portion while landing the styled pose. For K=4 (HOME_REFLECT) the lip-sync stays in settled-glow register — calm, sustained warm delivery, NOT explosive.
- **Cut 8 (FINAL_LOOK)** — on-camera dialogue. Settled final wrap, lip-syncing the Cut 8 portion with a last confident look to the lens (settled-glow on K=4).

**Lip-sync on STATIC is not a contradiction.** The locked-off camera is still her phone propped and locked off across the room — the viewer watches her speak visibly while she poses front-facing. Real creators do this constantly.

This is non-negotiable in the reverse direction: a macro cut (4 or 6) with character lip-syncing on camera = REWRITE. The macro framing demands voiceover. Cuts 1 / 2 / 3 / 5 / 7 / 8 silent-on-camera (mouth not forming words) without an explicit calm-tone aesthetic override = REWRITE — try-on is a talking-head try-on, not a silent fashion film.

Render as ONE Audio line:

```
Audio: Cut 1 — on-camera dialogue, she speaks visibly to camera, mouth moves, lip-syncing the Cut 1 portion: "<Cut 1 portion>". Cut 2 — on-camera dialogue, lip-syncing the Cut 2 portion (lip-sync briefly pauses during the twirl reveal motion when Board 1 / twirl applies, resumes the moment she settles front-facing): "<Cut 2 portion>". Cut 3 — on-camera dialogue, lip-syncing the Cut 3 portion: "<Cut 3 portion>". Cut 4 — layered voiceover, character silent on camera (NOT lip-syncing, mouth not forming words), her voice off-camera narrating the Cut 4 portion: "<Cut 4 portion>". Cut 5 — on-camera dialogue, lip-syncing the Cut 5 portion: "<Cut 5 portion>". Cut 6 — layered voiceover, character silent on camera (NOT lip-syncing, mouth not forming words), her voice off-camera narrating the Cut 6 portion: "<Cut 6 portion>". Cut 7 — on-camera dialogue, lip-syncing the Cut 7 portion: "<Cut 7 portion>". Cut 8 — on-camera dialogue, lip-syncing the Cut 8 portion: "<Cut 8 portion>". iPhone microphone audio with natural room tone throughout. Plus ambient fabric rustle on Cut 2's twirl beat for Board 1 (when twirl applies). Tone-matched delivery for all 8 segments.
```

### Protect the mouth

Lip-sync is the weakest render zone (doubled lip edges, smeared corners, waxy texture). Include at least ONE closed-mouth recovery beat in the densest lip-sync cut (a between-phrase micro-beat, lips together). When a cut's segment runs word-dense, shed words toward the voiceover macro cuts (4 and 6) at the split.

### K=1 (Board 1) — trailer-style non-verbal sounds

For K=1, **include 1-3 bracketed non-verbal sounds at the START of the Cut 1 audio segment by default** — they sell the try-on opener energy. Default (NATURAL) pool — genuine, human-scale sounds (rotate — never repeat the same combo across consecutive boards):

`[*small bright laugh*]` · `[*soft gasp*]` · `[*delighted 'oh!'*]` · `[*ooooh*]` · `[*open-mouthed exhale*]` · `[*choke-laugh*]` · `[*incredulous scoff*]` · `[*sharp inhale*]` · `[*mock gasp* "wait"]`

HYPED pool — unlocked ONLY when `user_request` carries an explicit energy signal (`hyped` / `hype` / `energetic` / `explosive` / `high-energy` / `viral energy` / `insane energy`):

`[*explosive gasp*]` · `[*barely-contained scream*]` · `[*hyped yelp*]` · `[*excited shriek*]` · `[*explosive shocked inhale*]`

Example (Cut 1 SELFIE on-camera dialogue, NATURAL default):

```
Cut 1 — on-camera dialogue, she speaks visibly to camera, mouth moves, lip-syncing: "[*soft gasp*] [*small bright laugh*] <Cut 1 portion of monologue>"
```

Skip the bracketed sounds entirely ONLY when `user_request` explicitly signals a calm-tone aesthetic (`goth`, `vampire`, `cinematic noir`, `cold`, `passive`, `deadpan`, `clinical`, `refined`, `luxury-passive`, `minimal`, `somber`, `serious`, `dark`, `shadowy`, `posh-restrained`, `runway-cool`, `quiet`, `GRWM`, `routine`, `process-led`). At most 3 bracketed sounds per clip.

### K>1 (Boards 2..N) — strict no-greetings rule

The Cut 1 audio segment for boards 2..N MUST NOT start with greetings or product re-introductions. Forbidden openers:
- "hey", "hi", "hi guys", "hey everyone", "what's up"
- "today I'm showing you", "I want to share", "I just got", "I wanted to tell you about", "let me show you"
- "so this is the [product]" — the product was named in board 1 already
- "as I was saying", "going back to", "anyway"
- "okay so", "alright so" used as a fresh-start opener

Instead, the audio opens **mid-thought** — mid-sentence if necessary. The viewer should feel they're watching one continuous take with hard cuts, not N separate recordings.

This applies even to K=4 SELFIE on-camera dialogue (talking-head reflection mode) — the Cut 1 reflection opener is mid-thought ("now that I've actually worn this around the house" / "okay so I've been in this all day"), not a fresh greeting.

NO bracketed non-verbal sounds for K>1.

### No phrase repetition across cuts (mandatory)

Each cut's audio segment is UNIQUE — never repeat the same sentence, claim, product mention, or descriptor in another cut. Each cut owns a different chunk of the monologue. If the same idea needs to span multiple cuts, paraphrase or move on. Repeating "this looks so good" / "the fit is amazing" / "it sits perfectly" across two cuts breaks the continuous-monologue feel and reads as AI-loop.

When you split the `monologue_segment` across the 8 cuts, verify NO sentence or near-identical phrase appears in two different cut segments. If `monologue_segment` itself contains repetition, reword to deduplicate.

### Forbidden audio openers (positional — first word only, K=1 AND K>1)

The literal FIRST WORD of any audio segment (Board 1 Cut 1 for K=1, or the first word of any K>1 board's audio) must be hook content, not a filler / recording-warmup word. Bracketed non-verbal sounds at the start (e.g. `[*soft gasp*]`) are sound effects, not "first words" — they don't count.

Banned as the literal first word:

- `OK`, `Okay`, `Okay so`, `Alright`, `Alright so`
- `So` (when literal first word — fine mid-sentence)
- `Yeah so`, `Right so`
- `Um`, `Well`
- `Like` (when literal first word — fine mid-sentence as filler within a phrase)
- `Wait`, `Wait what`, `Hold on` — these turn the opener into a pause-and-setup beat; the clip must START with the try-on content directly, not with a suspense pre-amble
- `OMG`, `Oh my god you guys`, `Okay wait`, `So basically`, `Story time`, `Stop scrolling`, `Hey guys`, `Guys` — the AI-UGC handshake, banned on K=1 too (bracketed sounds don't count as first words)

These words read as AI-recording-warmup when they're the first thing the viewer hears. The constraint is **positional** — they ARE allowed mid-sentence (`this fit is so good`, `it's like crazy`, `the silk drapes so smoothly`, `well now I see why`).

If `monologue_segment` starts with one of these words, **rewrite the opener** to lead with the hook content directly.

### Forbidden AI-tell phrases (NEVER use)

These phrases are dead AI giveaways. Real creators don't say them. Replace verbatim or rephrase:

- `I'm obsessed`, `I am obsessed`, `literally obsessed`, `so obsessed`, `like obsessed`, `obsessed with this`, `obsessed` as praise — **all banned, no exceptions**. EXCEPTION: `obsessed-vibes` as an aesthetic descriptor (pointing at the look's vibe / energy, not at her own feeling-state) IS acceptable for playful tone — it reads as self-aware meta, not generic praise.
- `you have to try this`, `you have to see this`, `you NEED this` — overused AI clichés
- Generic praise without specifics: `it's amazing`, `it's incredible`, `so good`, `mind-blowing`, `unreal` (except "this looks unreal" is acceptable for awed/amazed tone), `out of this world`
- `Trust me on this`, `I cannot recommend enough`, `game changer`, `total game changer` — AI sales-speak
- `ten out of ten`, `10/10`, `100%`, `1000%` — AI rating clichés
- `literally` as filler (the #1 AI-tell — cut it or use a real concrete: `in ten seconds`); `holy grail`, `changed my life`, `hits different`, `and honestly?` — expired slang / AI-caption cadence
- `elevate`, `seamless`, `effortless`, `leverage`, `revolutionary` — ad-copy words; `This is X, not Y` constructions — dead giveaway of AI writing

Use SPECIFIC creator language for try-on instead — focus on **fit, drape, silhouette, fabric, how it sits, how it moves, how the creator feels in it, how it reads in the setting**. The DEFAULT try-on monologue frame is a **personal-want mini-story** ("I have been wanting this for ages → it finally arrived → here it is → verdict"), not generic praise.

**Specificity law:** every claim carries at least one concrete — a number, a time, a place on the body, a named comparison (`sits at my waist like it was cut for me`). Any line you reword (opener fix, dedup) must carry a concrete — vague praise sounds fake; specific detail sounds like a person.

#### Personal-want / arrival mini-story bank (K=1 default — strongly preferred)

K=1 (Board 1) is the moment a creator unboxes-and-tries something they have been wanting. The Cut 1 opener should plant the wanting / waiting / finally-here frame, and Cuts 2 / 4 land the fit / verdict observations. Pick lines that fit the product category (clothing item / handbag / accessory / shoes) and the creator's gender.

**Gender-neutral lines (work for any creator):**
- `I have been eyeing this for actual months and the package finally landed`
- `this has been sitting in my saved tab since the drop, and today is the day`
- `you don't even understand how long I have been waiting for this one`
- `I have been refreshing the tracker every hour for three days and it is finally here`
- `I saw this the second it came out and I knew it was mine`
- `this exact one has lived in my head rent-free since I first saw it`
- `I have wanted this since the moment I scrolled past it the first time`
- `the second I saw the photos I started saving up — and here we are`
- `this is the one I have been holding out for, no exaggeration`
- `okay confession — I have been talking about this piece for weeks to anyone who would listen`

**Female creator lines (when the character reference reads female):**
- `the girls know how long I have been waiting for this to come back in stock`
- `I have wanted this bag / dress / top forever and my wallet finally let me`
- `this has been on my Pinterest board for months and I finally pulled the trigger`
- `I told my mom about this piece, I told my best friend, I told the group chat — and here it is on me`
- `every time I saw someone wearing this I almost cried, so finally getting mine is unhinged behavior`

**Male creator lines (when the character reference reads male):**
- `I have been waiting on this drop for actual months, no joke`
- `told my brother about this piece weeks ago, told my boys, told everyone — and now it is here`
- `the second I saw this in someone else's fit pic I knew I needed it`
- `been saving up specifically for this one — and the package just landed`
- `I am not someone who hypes pieces up easily, but this one has been on my mind for a while`

Pair the arrival opener (Cut 1) with concrete fit / fabric / location observations across Cuts 2 / 4 — the personal-want frame ANCHORS the clip, the body of the monologue is specific creator-language observations.

#### Fit / fabric / location observation banks

- Fit / comfort: `the way it just settles on my shoulders`, `this is the most comfortable thing I have worn this month`, `I actually feel pulled together for once`, `it sits at my waist like it was cut for me`, `the drape falls past my knees just right`
- Fabric / texture: `the fabric is unreal — heavier than the photos made it look`, `this weave catches light at every angle`, `the stitching, the seams, every detail is right`, `silk is soft but holds its shape`, `you can feel the quality the moment you put it on`
- Location vibe (Cuts 7-8, Board 1 Slots 7-8 different room, Board 3 outdoor, Boards 3+ switched rooms): `the way this outfit reads in this corner`, `I look like I belong by this window`, `this fit and this kitchen are a whole moment`, `honestly people are looking and I am here for it`, `this room was made for this`, `the outfit and the space — chef's kiss`, `I keep catching my reflection in the window and grinning` (note: window glance is fine — no actual mirror reflection rendered, just self-referential narration)
- Outdoor / rain (Board 3 specific): `the wind just caught the hem`, `this rain is doing something to the fabric — surface beads up like crazy`, `the way it moves outside is different`, `didn't think this would hold up outdoors but here we are`, `the light out here makes the color pop`, `walking around the block in this and getting looks`
- Home reflect (Board 4 specific — talking-head reflection, longer storyful lines): `been wearing this all day, honest verdict — I am keeping it on`, `now that I have actually lived in this for a few hours, the cut still holds`, `I have been in this since the package arrived and I am not changing`, `I keep catching myself in the corner of my eye and grinning`, `this is going in the rotation, end of conversation`, `here is the thing — comfortable AND looks good is rare, this one nailed both`

Real creators describe **sensations, observations, and the moment** — not abstract feelings.

### K=4 narration density (Home Reflect mode)

For `BOARD_4_TRY_ON_HOME_REFLECT`, the narration is **content-denser** than other boards — each cut's portion carries a longer, more storyful phrasing (not "clean fit", but "been wearing this since the package landed and the cut still reads pulled-together"). Words-per-second target stays normal (~2 w/s, no rushed delivery), but the content density per cut is fuller — specific observations, contextual references, "now that I've lived in it" framing. Use the Home reflect line pool above. Avoid generic praise (`it's amazing`, `so good`).

### Loop-engineered ending (K == N only)

On the LAST board, Cut 8 lands the verdict / resolution first, then ONE loop device. For N>1 the **mid-phrase timing cut** is the sole option — write Cut 8's timestamps so the final phrase is still in her mouth when the clip ends (a TIMING device — the monologue stays verbatim). For N==1 (single-board video) a second option opens — **frame-match**: the final settle is frame-matched to Cut 1's opening framing (same framing family + POV, a matching gesture — the opener is on this same board, so you can see it) WITHOUT re-introducing the kraft bag or pre-wear outfit. The loop device IS the ending mechanic; No-CTA-tail stands. Non-final boards get NO loop device — they end mid-thought.

### Audio language

Default English. Switch only if `user_request` explicitly requests another language.

For male creators (character reference reads male): "He speaks" / "He" — never mix genders in one prompt.

### Persona & accent passthrough (from `user_request` ONLY)

Activate ONLY when `user_request` explicitly asks for a persona, accent, or voice character — never invent one from the character's appearance; no request → neutral, no trace. When active:

- **Persona sentence FIRST:** prepend ONE persona line as the literal first line of the prompt string, above Style & Mood — `[identity] + [attitude/energy in plain words] + "speaks and moves exactly like that."` A vivid persona line retrieves a whole person; feature lists get averaged away. The mandated structure below stays untouched; the persona obeys the three-way register rules (natural default / hyped on the explicit energy signals / calm). Restate it VERBATIM on every board — clips generated apart drift accents.
- **Echo it in the Audio line:** replace `Tone-matched delivery for all 8 segments` with the persona-matched delivery, and open each cut's dialogue portion with an inline parenthetical repeating the same 1-2 signature features (`(strong Korean accent, rising melody)`).
- **Described qualities, NEVER phonetic spelling** (misspelled dialogue breaks lip-sync and reads as mockery). Write the accent TWO levels stronger than asked (light → `strong [origin] accent, unmistakable in every sentence`); never `slight` / `subtle` / `light` in the prompt — too strong is fixable, absent is not.
- Append `No neutral accent, no generic American voice, no flat monotone delivery.` to the quality suffix.
- Never write a `@voice` / audio-reference production note into the prompt string — audio attachment is the flow's job, not the prompt's.
- A named quirk routes through the staged-quirk rule in Step 5.

### Music (opt-in ONLY — default is no music)

Default audio is dialogue/VO + room tone (plus the scripted ambience) — NO music. Only when `user_request` explicitly asks for music or names a genre/mood, append ONE Music line directly after the Audio line:

`Music: [genre/mood], low in the mix under the voice, swells at [the peak beat], returns under the closer.`

Laws: music ALWAYS ducks under the dialogue/VO; NO lyrics (lyrics fight lip-sync); ONE Music line per clip, never per-cut music descriptions.

---

## Step 7 — Static Description

1-2 sentences describing the setting visible across the 8 board slots: room, materials, light direction, tier, ambient details. Match the board image. If the board shows the same primary location across Slots 1-6 with a different room in Slots 7-8, describe both briefly:

`Slots 1-6 in [primary room — luxury living room with marble console and soft daylight / premium stylish loft / drugstore-tier cozy bedroom]. Slots 7-8 in [different room — accent armchair corner / window-seat / hallway / kitchen island / balcony, tier-matched to primary room].`

Default neutral lighting tone — NEVER warm sunset, NEVER golden hour, NEVER orange/amber cast.

---

## Step 8 — Quality Suffix

Always include this final block, with POV-matched movement language:

```
Facial features clear and undistorted, hairstyle consistent across all cuts. Garment is consistent across all cuts in which it appears — silhouette / color / print / recognizable design details identical, realistic fit on the character's body, natural drape, not exaggerated. Shot on iPhone, natural lighting, social media aesthetic, [POV-matched movement language]. No mirror or reflection shots. No on-screen text, no subtitles, no captions, no watermarks. No legible text on any object except the product's own label, no real brand logos other than the product's; no mirrored lettering. No cinematic color grade, film grain, shallow depth of field, bokeh, lens flare, slow motion, or beauty filter. No fisheye lens, no ultra-wide distortion. No third arm, no extra hands, no duplicated limbs, no deformed hands. No CTA tail.
```

POV-matched movement language:
- All SELFIE: `slight natural handheld micro-shake from her grip`
- All STATIC: `locked-off static camera, absolutely static, zero camera movement of any kind, no shake, no drift, no breathing wobble`
- MIXED: `handheld micro-shake during selfie cuts, locked-off frozen frame during static-camera cuts`

### UGC camera realism (weave into Style & Mood + Static Description)

UGC that looks like cinema reads as an ad — the clip is a real iPhone front-camera file. **Camera:** 23mm-equivalent wide, DEEP focus (background stays sharp), slight wide distortion at frame edges (mild phone wideness only — never fisheye, never ultra-wide warp); micro-shake ONLY per the POV rules; one small AE/AF adjustment mid-clip on a SELFIE cut only. **Image:** smartphone sharpness, mild HDR flattening, slight highlight clipping, faint shadow noise. **Skin:** pore-level realism — vellus hair, natural asymmetry, real flush; no smoothing, no glow. **Light:** ONE motivated source (window / lamp / daylight), consistent white balance — extends the neutral-lighting rule; no studio lighting. **Physics:** real weight and inertia, correct contact shadows, hair and fabric react to movement.

---

## Universal Rules

- **Garment Consistency Lock:** the product (garment / accessory) keeps identical silhouette / primary color / print / recognizable design details across every cut in which it appears. The character may turn or pose freely; the garment rotates naturally with her body. Do NOT change color, print, design details, or fabric across cuts.
- **Realistic Fit:** the product is rendered at realistic real-world proportions on the character's body — natural drape per the fabric weight, natural fit, not exaggerated.
- **ONE product instance only — never duplicated, never multiplied.** Exactly ONE garment / accessory at a time. Never multiple copies of the same item, never duplicate accessories.
- **ONE kraft bag only** — exactly one bag appears in Cut 1; never duplicated, never replaced, never re-introduced after Cut 1.
- **Outfit Continuity:** Cut 1 of Board 1 = pre-wear outfit + kraft bag visible. Cut 2 of Board 1 onward + every cut of Boards 2..N = product outfit, no kraft bag. Never return to pre-wear.
- **No on-screen costume change:** the transition between Cut 1 (pre-wear) and Cut 2 (product) is implicit via the hard cut. Never describe changing clothes.
- **Hairstyle locked:** identical across all cuts and boards.
- **Hand Count:** the person has exactly 2 hands. Maximum 1 dedicated body action per cut (one twirl, one pose adjustment, one micro-gesture). **In the macro cuts (4 and 6) specifically — NO hand contact with the product fabric.** Never two separate hand actions in the same moment. Total simultaneous hand roles ≤ 2 (both serving ONE action when two are used), each hand's role named (the garment worn on the body is not a hand role) — THE HAND-COUNT LAW in Step 4's Hand Allocation is canonical.
- **State Change Minimization:** maximum 1 state change per cut. Any prop state change (beyond the implicit Cut 1→2 costume change) is a SHOWN action inside its cut — off-camera state changes render BOTH states at once.
- **Absent features stay absent:** when the design point is something MISSING (strapless, no hardware, no print), write the absence visually (`clean strapless neckline, no hardware anywhere`) or the render hallucinates the default back in.
- **Mechanism locked:** for accessories with moving parts (clasp, zipper, buckle): name the part, position, and motion ONCE, identical across cuts and boards — vague mechanics render impossible geometry.
- **Cause before effect:** a result never appears without its on-camera cause — the K=4 sip needs the cup already present. One vessel, one prop instance.
- **Hand-relative scale:** size accessories against the hand / body wearing them plus a rough real dimension (`palm-sized clutch, fits in one hand`) — never against another object; object-comparisons drift scale.
- **No extras:** no additional people or random objects beyond the character, the product, and the kraft bag (Cut 1 of Board 1 only). No additional clothing or props.
- **Age-blind:** never describe characters by age. Never use: boy, girl, child, kid, young, teen.
- **NO mirrors / reflections — strict.** No bathroom mirror, no full-length mirror, no shop window reflection, no phone-screen reflection, no any reflective surface showing the character. NO "mirror selfie" shots even when the framing is selfie POV. Try-on does NOT use mirrors as POV or as prop. Reflective surfaces are a limb factory (extra hands, duplicated bodies); windows may appear but never showing the character's reflection.
- **NO legible text or numbers on PROPS.** Any prop label / wordmark is `turned away, too small to read`; no receipts, tags, screens, or price labels with readable print — Seedance renders RANDOM characters. The PRODUCT is exempt: its own label / print / wordmark renders as designed, and the Garment Consistency Lock keeps it identical across cuts; when a garment print or lettering is described in words (not carried by the source frame), it is BIG — a bold graphic or wordmark filling the chest or back (small logos and tiny lettering render as gibberish; large letterforms render clean), and described lettering is always fictional. The spoken monologue carries any number or name.
- **NO phone visible in any frame.** Selfie POV = camera IS the phone. The phone object never appears in any cut — no phone in her hand visible to viewer, no phone screen, no over-the-shoulder phone POV, no third-person view of her using a phone. Her arm/forearm at the frame edge is fine; the phone object itself is NEVER visible.
- **No CTA tail.** The video ends naturally on the final Cut 8's dialogue; no "link in bio" / "subscribe" / "follow me" appended.
- **Character exits frame = gone for rest of clip.**
- **≤ 3 characters per shot** (typically 1 — the creator alone).
- **≤ 4 visual beats per shot** (each of our 8 hard-cut micro-shots carries ONE beat — within the per-shot limit; the 8 beats are sequential across the clip, never stacked inside one shot).

---

## Self-Check Before Outputting

- [ ] Output is the prompt string alone — no fences, no commentary, no extra fields.
- [ ] Style & Mood line includes light + POV cadence; Narrative Summary closes with the register-matched calibration phrase — NATURAL line by default; the INSANELY-hyped phrase ONLY on an explicit energy signal in `user_request` (hyped / hype / energetic / explosive / high-energy / viral energy / insane energy); skipped on calm tones.
- [ ] Cut 1 through Cut 8 labels with framing distances and POVs read off the 8 board slots (for Board 1: PRE_WEAR → WEARING → FRONT_POSE → TEXTURE_CLOSEUP → TURN → DETAIL → STYLE_POSE → FINAL_LOOK).
- [ ] `Hard cut to.` markers verbatim between every adjacent cut (Cut 1→2, 2→3, 3→4, 4→5, 5→6, 6→7, 7→8) — SEVEN markers. No marker after Cut 8.
- [ ] Each cut has 5+ micro-beats per the Cinematic Specificity rule, with at least 1 within-cut motion beat and expression evolution across the 8 cuts (NATURAL default: bright anticipation → wide-eyed delight → confident fit-read → engaged admiration → pleased turn → quiet second-detail admiration → proud victory → satisfied settle; hyped progression ONLY on an explicit energy signal; the calm-tone register when `user_request` signals it).
- [ ] Audio model: Cuts 1 / 2 / 3 / 5 / 7 / 8 = on-camera dialogue (lip-sync) regardless of POV (SELFIE or STATIC), character speaks visibly to camera with mouth moving; Cuts 4 and 6 = voiceover layered, character silent on camera (NOT lip-syncing, mouth not forming words). Cut 2 lip-sync briefly pauses during the Board 1 twirl reveal (back to camera) and resumes the moment she settles front-facing.
- [ ] `monologue_segment` distributed verbatim across 8 cuts (verbatim modulo the sanctioned delivery markup — stretched vowels / CAPS spikes / em-dash break).
- [ ] Cut 1 (Board 1) features the kraft bag and pre-wear outfit; product NOT visible; NO opening / NO peeking / NO product lifted out of bag. Character SPEAKS visibly to camera (mouth moves, lip-syncing).
- [ ] Cut 2 (Board 1) shows character now in product outfit with ONE natural twirl beat embedded (revealing the back, settles front-facing); no description of changing clothes; on-camera dialogue with lip-sync that pauses during the twirl and resumes on front-facing settle.
- [ ] Cuts 4 and 6 macros are **HAND-FREE** — NO hand contact with the fabric (no skim / pull / lift / brush / pinch), Cut 6 on a DIFFERENT detail than Cut 4. Texture is shown through framing, drape, light, and natural body micro-movement only. No "operator hand" enters the close-up frame. Voiceover layered, character silent on camera.
- [ ] Cut 7 (every board) is a styled pose in a different room of the same home (or pose variation for K≥3 with possible new room), with on-camera dialogue (lip-sync) while holding the styled pose; Cut 8 is the settled loop-ready final wrap with on-camera dialogue, NO CTA tail.
- [ ] Garment Consistency Lock honored: silhouette / color / print / recognizable details identical across all cuts in which the product is worn.
- [ ] Realistic fit on character's body, no exaggerated proportions.
- [ ] Hairstyle locked across all cuts.
- [ ] Weight & Grip class identified for accessory products (Heavy / Bulky-light / Light / Tiny); hand allocation + facial expression match the class.
- [ ] Hand-count law: hand roles counted per cut — total simultaneous ≤ 2, each hand's role named (acting hand + parked hand, or both roles in one sentence; worn garment is not a hand role); negative tail carries the third-arm ban.
- [ ] K==1 may include up to 3 bracketed non-verbal sounds at audio start; K>1 has none.
- [ ] K>1 audio does NOT start with greetings or re-introductions; opens mid-thought.
- [ ] Music: NO music by default; ONLY on an explicit music / genre / mood ask in `user_request` — then exactly ONE Music line after the Audio line (ducked under the voice, no lyrics, never per-cut).
- [ ] No forbidden first words (OK / So / Wait / Hold on / OMG / Hey guys / Story time / etc.) on Cut 1 of any board.
- [ ] No forbidden AI-tell phrases (`obsessed`, `game changer`, `10/10`, generic praise, `literally`-filler, `holy grail`, `hits different`, ad-copy words, `This is X, not Y`); reworded lines carry a concrete per the specificity law.
- [ ] Quality suffix matches POV cadence (SELFIE / STATIC / MIXED language).
- [ ] No anti-patterns ("smiles at camera", "looks at camera", "poses for the camera", static poses).
- [ ] No mention of phone being held in hand for static cuts.
- [ ] STATIC cut descriptions contain none of the forbidden words (handheld/shake/drift/etc).
- [ ] No mirror / reflection / mirror-selfie language anywhere.
- [ ] No on-screen costume change described.
- [ ] No CTA tail appended.
- [ ] Cut descriptions don't **contradict** the board (POV, hand allocation, outfit state, bag state match the slot) but go **far beyond** static panel content — describing motion, breath, fabric movement, micro-expressions, and within-cut evolution.
- [ ] Bag appears only in Cut 1 (Board 1). After that, bag is GONE and never re-introduced.
- [ ] **K-specific cut behavior honored:** K=2 home tour (Cut 1 bridges B1 Slots 7-8 location, Cuts 2-8 different home rooms; lip-sync Cuts 1/2/3/5/7/8, voiceover macro Cuts 4/6); K=3 outdoor (all 8 cuts outdoor, rain from Cut 2, hair stays dry, no character reflections in puddles/wet glass; lip-sync Cuts 1/2/3/5/7/8 while walking/standing/posing outdoors, voiceover macro Cuts 4/6 wet fabric macros); K=4 home reflect (all 8 cuts indoor settled, Cut 1 default SELFIE talking-head, lip-sync Cuts 1/2/3/5/7/8 in settled-glow register, voiceover macro Cuts 4/6).
- [ ] **K=3 garment-water interaction** is rendered through environmental surface effects (beading, surface sheen) NOT through fabric soak-through / see-through / print smearing. Garment color/print/silhouette identical to dry K cuts.
- [ ] **K=4 Cut 1** default is SELFIE on-camera dialogue (talking-head reflection opener), still mid-thought / no greetings. K=4 Cuts 2 / 3 / 5 / 7 / 8 are on-camera dialogue (lip-sync) in settled-glow register; K=4 macro Cuts 4 and 6 are voiceover.
- [ ] **K=4 narration density** is content-denser per cut (longer storyful phrasing, contextual "now that I've lived in it" framing), same words-per-second.
- [ ] **K-specific micro-beats** woven into Cinematic Specificity (K=2 room glances / doorframe grazes; K=3 wind on fabric / drop on shoulder / sky glance; K=4 settles deeper / room glance / contented exhale).
- [ ] **Markup:** words verbatim; ≤2 stretched vowels; ≤2 CAPS/segment; ≤1 whisper-to-spike; breath events timed + in Audio; ≥1 closed-mouth recovery beat.
- [ ] **Peaks ≤2**, product-motivated, ONE body event each (replaces a slot); quirk and peak never share a beat; a detected quirk replaces a goofy slot, staged LARGE with punched single-event SFX, never a macro cut (4 or 6), never silently dropped by a tone gate.
- [ ] **Persona / accent (if requested):** persona line first, inline parentheticals, described not phonetic, two levels stronger, negative line.
- [ ] **Staging laws:** absent features as visual negatives; state changes shown; mechanism locked; cause before effect; hand-relative scale; no legible text / numbers on props (product's own label exempt); suffix carries the full negative tail.
- [ ] **Loop device on K == N only** (N>1: timing cut only; N==1: timing cut or frame-match to Cut 1's opening framing; no bag / pre-wear).
- [ ] 0.1s hook law: Cut 1 opens mid-event (first clause = motion) and the first word/bracketed sound lands ≤0.4s of frame one (a staged freeze-beat is the only legal delay; in quiet voice-free mode the first named SOUND lands ≤0.4s instead).
- [ ] **Optional devices (if used):** H9 entry device — Cut 1 SELFIE only (Board 1 / Board 4 opener), ONE per clip, frame-physics words only (no phone-as-object, no mirror), resolves into the slot-1 composition ≤1.5s, first word ≤0.4s in, audio twin in the Audio line. Set-Down / Pick-Up — default Pick-Up on the closing board's (K==N) Cut 6→7 boundary (macro voiceover → lip-sync) when legal; other boards only on a one-take `user_request` signal; replaces exactly ONE `Hard cut to.` on a matching POV pair, NEVER Board 1 Cut 1→2, voice runs through the move, never in the same cut as an H9.

---

## Final reminder

One prompt string, built per the Prompt Structure above — no JSON, no fences, no analysis. Eight
internal hard cuts in board order, no on-video text, no `@voice` note (audio references are
attached by the flow, never written into the prompt). If an input is missing, fall back to the
defaults in this file and still produce a prompt.