Set prompt video lengkap

---
=== LOCK (same in Segment 1, 2, 3 - do not change) ===
Avatar: [AVATAR_DESCRIPTION - e.g. Indonesian male, round thin-frame glasses, dark wavy hair, thin goatee, warm brown skin], SEATED at a desk, plain solid-black crew-neck t-shirt (no logo/print), framing chest-to-top-of-head, never standing or walking.
Environment: Indonesian home office at night. Warm tungsten lamp rear-right, bookshelf behind. Dark warm amber mood, deep shadows, no cool/blue tones. Same room in every segment.
Format: 9:16 vertical, exactly 10 seconds, 15% safe zone top and bottom, no letterbox.
Floating text: exists in 3D space inside the scene (NOT a flat 2D overlay) - hangs from a 1px warm-white (#FFF8F0) hairline thread with subtle organic sway; depth layers foreground/midground/background.
---

Generate a 10-second vertical video clip (9:16, 1080x1920px). This is Segment 1 of 3.

CAMERA: Slow push-in from medium shot. Camera drifts 15% closer over 8 seconds. Shallow depth of field. Warm tungsten backlight from rear-right. Soft key light from left side.

AVATAR ACTION: [AVATAR_DESCRIPTION — mis. "Indonesian male, round thin-frame metal glasses, dark wavy hair, solid black t-shirt, SEATED"] Opens with 0.5 seconds of direct eye contact with camera — calm, confident — then delivers monologue. One eyebrow raise on the hook word.

MONOLOGUE (spoken in Indonesian):
"[NAMA_TOOL] punya fitur yang kebanyakan orang nggak tau cara pakainya.
Dan itu yang bikin output mereka biasa-biasa aja.
Gue kasih tau cara pakainya yang bener — dan hasilnya beda banget."

FLOATING TEXT:
- At 0:02 — display "[HOOK_WORD]" in bold terracotta (#C4703A), 72px, centered, bottom third of frame, hanging from 1px hairline thread (#FFF8F0), thread swaying subtly
- At 0:06 — display "[3–4 WORD BENEFIT]" in off-white (#FFF8F0), 44px, midground depth, below terracotta word, separate thread
- Text enter: drop from above, decelerate into position. Hold 3 seconds. Exit: retract upward.

ENVIRONMENT: Indonesian home office interior, night. Warm tungsten lamp rear-right. Bookshelf in background. Dark warm amber mood — deep shadows, no cool tones.

MOOD: Calm, authoritative. The presenter knows something the viewer doesn't — not arrogant, just confident.

---
=== LOCK (same in Segment 1, 2, 3 - do not change) ===
Avatar: [AVATAR_DESCRIPTION - e.g. Indonesian male, round thin-frame glasses, dark wavy hair, thin goatee, warm brown skin], SEATED at a desk, plain solid-black crew-neck t-shirt (no logo/print), framing chest-to-top-of-head, never standing or walking.
Environment: Indonesian home office at night. Warm tungsten lamp rear-right, bookshelf behind. Dark warm amber mood, deep shadows, no cool/blue tones. Same room in every segment.
Format: 9:16 vertical, exactly 10 seconds, 15% safe zone top and bottom, no letterbox.
Floating text: exists in 3D space inside the scene (NOT a flat 2D overlay) - hangs from a 1px warm-white (#FFF8F0) hairline thread with subtle organic sway; depth layers foreground/midground/background.
---

Generate a 10-second vertical video clip (9:16, 1080x1920px). This is Segment 2 of 3. Room, avatar, outfit, and floating text system identical to Segment 1.

CAMERA: Static medium shot with very subtle lateral drift (left to right, max 8% frame shift over 10 seconds). Subject centered. No cut — single continuous shot.

AVATAR ACTION: Same avatar as Segment 1 — SEATED, black t-shirt, round metal glasses. Slightly more animated than S1 — natural hand gestures near chest level to illustrate the steps. Can count steps on fingers.

MONOLOGUE (spoken in Indonesian):
"Pertama, [LANGKAH_1 — satu kalimat singkat].
Kedua, [LANGKAH_2 — satu kalimat singkat].
Ketiga, [LANGKAH_3 — satu kalimat singkat].
Ini yang jarang orang lakuin — tapi ini yang bikin hasilnya beda."

FLOATING TEXT:
- At 0:01 — display "1. [KATA_KUNCI_LANGKAH_1]" in off-white (#FFF8F0), 48px, upper-left quadrant, midground depth, with 1px hairline thread (#FFF8F0)
- At 0:04 — display "2. [KATA_KUNCI_LANGKAH_2]" in off-white (#FFF8F0), 48px, center-left, midground, thread attached
- At 0:07 — display "3. [KATA_KUNCI_LANGKAH_3]" in terracotta (#C4703A), 52px, center-left, midground, thread attached — this is the key step, highlight it
- Each text element slides in from the left, 0.25s ease-out. Remains on screen until end of segment.

ENVIRONMENT: Identical to Segment 1 — same Indonesian home office, same tungsten lamp rear-right, same bookshelf, same dark warm amber mood. Do not change anything about the room.

MOOD: Instructional but energetic. The viewer is learning — keep it moving.

---
=== LOCK (same in Segment 1, 2, 3 - do not change) ===
Avatar: [AVATAR_DESCRIPTION - e.g. Indonesian male, round thin-frame glasses, dark wavy hair, thin goatee, warm brown skin], SEATED at a desk, plain solid-black crew-neck t-shirt (no logo/print), framing chest-to-top-of-head, never standing or walking.
Environment: Indonesian home office at night. Warm tungsten lamp rear-right, bookshelf behind. Dark warm amber mood, deep shadows, no cool/blue tones. Same room in every segment.
Format: 9:16 vertical, exactly 10 seconds, 15% safe zone top and bottom, no letterbox.
Floating text: exists in 3D space inside the scene (NOT a flat 2D overlay) - hangs from a 1px warm-white (#FFF8F0) hairline thread with subtle organic sway; depth layers foreground/midground/background.
---

Generate a 10-second vertical video clip (9:16, 1080x1920px). This is Segment 3 of 3 — the payoff and call to action. Room, avatar, and text system identical to Segments 1 and 2.

CAMERA: Stable, centered medium shot. Camera holds still for the first 7 seconds — stillness signals confidence in the solution. At second 7, very slow zoom-in (5% closer) as CTA text appears.

AVATAR ACTION: Same avatar — SEATED, black t-shirt, round metal glasses. Delivers final line with a calm, knowing energy. Direct eye contact with camera throughout. At the CTA line, a slight lean forward — closing the distance with the viewer.

MONOLOGUE (spoken in Indonesian):
"Kalau kamu lakuin tiga langkah ini dengan urutan yang bener, hasilnya langsung kerasa.
Gue udah coba di [JUMLAH] video dan konsistensinya jauh lebih bagus.
Simpan video ini — dan follow buat tutorial [NAMA_TOOL] berikutnya."

FLOATING TEXT:
- At 0:02 — display "[HASIL_UTAMA — max 4 words, all caps]" in off-white (#FFF8F0), 52px, upper third, centered, 1px hairline thread (#FFF8F0)
- At 0:07 — display "SIMPAN" on the left side, "KOMEN" in terracotta (#C4703A) center, "FOLLOW" on the right side — all 48px, lower third, each with their own 1px hairline thread, entering together in rapid succession
- CTA text hold for 3 seconds until segment ends.

ENVIRONMENT: Same as Segments 1 and 2. Optional: very subtle warmth increase in color grade — like the room settling. Do not change the room itself.

MOOD: Empowering and clear. Viewer leaves knowing exactly what to do next.