Promethee Turns Work into a Shared Game

A 105-second cinematic launch film makes an everyday activity feel like entry into a larger world. It opens with a second-person question, introduces a place and protagonist, then demonstrates the product through brief interface inserts before closing on a human connection. The promise arrives in the first 8 seconds: the viewer's ordinary effort can become visible progress.

Why it works

• The opening turns a familiar gaming behavior into a personal question by 0:08, over a silhouetted figure against a bright amber doorway.
• Live-action anticipation (0:00–0:30) gives the interface reveals at 0:30–1:19 a reason to matter; screens are intercut with hands and the protagonist.
• The ending widens the reward from private progress to other people at 1:21–1:29, then holds a nearly empty frame and final line through 1:44.

Hook formula: question plus visual pattern interrupt: “You've spent {{TIME}} improving {{FICTIONAL_PROXY}}. What if {{REAL_PERSON}} could grow instead?”

Why it holds: The first cut is at 0.27 s, but the shot then lasts more than 8 s. The curiosity comes from the gap between epic game imagery and an everyday second-person question. The specific product name waits until about 0:10.

Retention map: visual and narrative re-hooks at 0:09 (face/world), 0:30 (product reveal), 0:47 (quest), 1:06 (progress card), and 1:21 (guild). Feature payoff runs 0:30–1:19; emotional payoff begins 1:21. The post caption carries the availability CTA; the film ends on a hard, quiet statement rather than looping.

Mix: Cinematic live-action footage at the opening and payoff; screen recordings and designed UI composites through the middle; sparse black frames at transitions and the ending. The percentages in section 2 are visual estimates, not machine measurements.

Rhythm: The edit begins with a 0.27 s black prelude, then an 8.11 s silhouette hold. Median shot is only 1.97 s because demonstrations and human reactions cluster into 0.6–4.7 s shots. A long product shot at 0:57–1:15 has five internal changes; the final 19.12 s shot quiets down after a few early changes.

Rules this editor follows (inferred): Keep the opening and final statement spacious. Put quick hand, face, and interface inserts around important actions. Let longer app screens earn their duration through changing overlays. Cut most often around the mechanism reveal, not evenly throughout.

Full shot list

• Look: Cinematic realism: near-black #080808, dark brown #2A160C, amber #F1A01A, warm stone #8B7D70. Muted skin tones, black wardrobe, monumental concrete and warm side/backlight. Product UI switches to off-white #F5F5F2, charcoal #17191B, violet #6F2A75, and muted coral #DA705C. Approximate colours from the supplied frames.
• Captions: Small white sentence-case subtitle, closest free match Inter Regular, roughly 18–22 px on 1280×720, horizontally centered in the bottom black bar around y=670. One sentence or short clause per line; no box, outline, bounce, or highlighted words. Change at phrase boundaries rather than word by word.
• Graphics and overlays: Subtle game HUD panels: dark translucent rounded rectangles, small profile avatars, levels, rank/XP counters, compact toast notifications. UI is readable as a category of action even when its tiny body copy is not. A white end line sits centered on black.
• B-roll: Purpose-shot human action: walking, doors, sitting, hands on laptop, eyes, trackpad. Alternate wide architecture with intimate close-ups. Most inserts last under 4 s.
• Transitions: Mostly straight cuts; black frames and holds provide chapter changes. No conspicuous wipe/glitch system.
• Branding: The product name arrives in narration near 0:10 and interface around 0:30. No large logo end card in the supplied sheet.

Measured: 108 recognized words; 73 wpm over the whole 104.6 s and 111 wpm while speaking. Sixteen pauses of at least 0.25 s; median 1.57 s, longest 7.42 s before the app demonstration. Pitch median 138 Hz, 2.9-semitone movement SD; endings: 4 fall, 3 rise, 6 level. The transcription around product/UI names and 1:00–1:19 is noisy, so write new lines from the visible mechanism rather than copying it.

Described: A calm, intimate, medium-low narrator addressing one viewer. Light wonder in the opening question, confident restraint for the product explanation, an ordinary conversational register for momentary first-person reactions (inferred). Pauses give visual reveals room; the final belonging line lands slowly.

Voice direction: Use an original adult voice with a warm medium-low register and understated cinematic delivery. Speak about 105–115 words per active minute. Let the first question rise slightly, give the product name a clean beat, then use short declarative clauses for the proof. Leave 1–3 seconds around major visual reveals and a much longer space before the final thought. Avoid a sales read.

• Music: A continuous tonal, cinematic bed is measured between spoken words. The last section suggests roughly 121 BPM (uncertain tempo estimate). Build from sparse, suspenseful ambience into a restrained pulse during the interface section, then thin it for the final black frame (energy curve inferred from pacing and measured bed, not identified instruments).
• Sound effects: The cut-accent measurement finds no consistent hit on every cut. If adding sounds, use a few soft door, keyboard, notification, and interface ticks tied to visible actions; keep them secondary.
• Silence and space: The narrator leaves many substantial pauses, including a 7.42 s gap before the app explanation. After the last word, the tail falls to about −32.8 dBFS.
• Mix targets: Reference is −17.2 LUFS integrated with −0.2 dBFS true peak. For a new delivery, aim around −16 to −14 LUFS and ≤−1 dBTP to preserve headroom; keep speech intelligible over the score. The automatic voice/bed level separation is unreliable here, so mix by listening.

• Title: Product introduction + evocative mechanism: “Introducing {{PRODUCT}}: turn {{ACTIVITY}} into {{TRANSFORMATION}}.”
• Cover / thumbnail: Anonymous black silhouette centered before a tall amber portal, with broad black bars and one small white subtitle. The story image carries the click; no headline slab.
• Description: A one-sentence product promise, availability/status, and link. The source post uses that pattern. No reliable hashtag or pinned-comment pattern is supplied.
• Audience signals: About 857,000 views, 3,841 likes, 246 comments, and 288 reposts at capture. No most-replayed data or top comments were supplied, so audience response to individual beats cannot be determined.

Aim for about 105–115 spoken words across 105 seconds, with large spaces; the counts below sum to about 110. These are fresh template lines, not a rewrite of the source.

[0:00–0:09 · HOOK · ~17 words]
You know the feeling of improving {{PROXY}}. What if {{ACTIVITY}} showed your growth just as clearly?

[0:09–0:20 · WORLD · ~14 words]
Welcome to {{PRODUCT}}. A place where {{AUDIENCE}} are already {{ACTION}} together.

[0:20–0:30 · INVITATION · ~5 words]
Now step in and begin.

[0:30–0:57 · MECHANISM · ~30 words]
When you {{INPUT}}, {{PRODUCT}} starts {{MODE}}. It notices {{EFFORT}} and turns it into {{MEASURE}}. {{FEATURE_1}} makes the next step visible. {{FEATURE_2}} gives you a reason to continue.

[0:57–1:20 · PROGRESS · ~27 words]
As the work adds up, you can see {{RESULT}}. Save the session. Share what changed. Give someone else a glimpse of what is possible.

[1:20–1:44 · SOCIAL PAYOFF · ~17 words]
Then you see {{PEER_SIGNAL}}. You look up and discover {{COMMUNITY_REVEAL}}.

Script prompt

Write an original 105-second launch-film script for {{PRODUCT}}, which helps {{AUDIENCE}} do {{ACTIVITY}} and achieve {{TRANSFORMATION}}. Follow the six timed beats in this blueprint. Use about 110 spoken words total, second person at the opening, a single question by 0:08, a world invitation at 0:09–0:30, concrete product actions at 0:30–1:20, and a human/community payoff at 1:20–1:44. Give each beat a voice line, visual action, UI evidence, and planned pause. Use short clauses and restrained cinematic language. Do not borrow the source's character-leveling wording, product name, lines, cast, or exact interface.

Voice prompt

Original adult narrator, warm medium-low register, intimate and measured. About 110 wpm while speaking. Curious rise on the opening question, clean pauses before reveals, grounded statements for the demonstration, soft assured ending. No imitation of an identifiable person.

Visual prompts

• World shot: Original adult worker seen from behind approaching a luminous architectural entrance, warm amber backlight, deep black foreground, wide 16:9 frame, cinematic realism, slow restrained camera move; no borrowed logos or distinctive characters.
• Human insert: Close-up of an anonymous adult worker's eyes, hands, or trackpad in a quiet stone-and-wood workspace, warm side light, muted neutral grade, shallow focus, deliberate natural motion.
• Product insert: Faithful screen capture of {{PRODUCT}} showing {{FEATURE}} and its actual state change; overlay a subtle custom HUD for {{MEASURE}} without hiding the real interface.
• Final group: Three distinct working adults in a shared space, eyeline connection across cuts, warm practical lights and charcoal wardrobe, intimate close-ups, no celebrity likeness.

Caption spec: Inter Regular, about 20 px at 1280×720, white #F6F6F6, centered in a 65–90 px black bottom band; one short clause per card, hard replacement at phrase change, no highlight.

Music prompt: Cinematic ambient pulse, low tension into hopeful breadth, about 120 BPM or a half-time feel, sparse opening, restrained rhythmic middle, long quiet decay at the end; instrumental and licensed/original.

1. Write the six beats and record the original narrator first. Leave the gaps in the timeline rather than filling every second with speech.
2. Place the opening black frame, amber silhouette, door, face, and shared-world footage across 0:00–0:30.
3. Capture real {{PRODUCT}} states for {{INPUT}}, {{MODE}}, {{MEASURE}}, {{FEATURE_1}}, {{FEATURE_2}}, and {{RESULT}}. Add simple HUD elements only where they clarify the state.
4. Assemble the interface proof from 0:30–1:20, with quick human inserts around button presses and reactions. Keep the long progress card changing internally.
5. Cut to peers and a black final hold after 1:20. Place one restrained white closing line at about 1:35.
6. Add small lower-bar subtitles, then the score, then only a few action-based sounds. Mix speech for clarity and check the full film at normal phone volume.

• A relatable question and cinematic visual promise are clear by 0:08; product is named around 0:10.
• The actual product mechanism is visible by 0:30, with each major claim backed by a real screen state.
• Target 30–33 shots, roughly 17–19 cuts/min, and about 25–27 total picture changes/min; opening and ending deliberately hold longer.
• Narration is about 105–115 words at roughly 105–115 wpm while speaking, with pauses before product and social reveals.
• Subtitles are small, white, centered, and legible in the lower black band; UI labels remain distinct from them.
• Final social beat gets a quiet black hold; the film ends with a statement, while the post caption supplies any link/availability CTA.
• Final sound is intelligible and peaks ≤−1 dBTP.

• The source signature (don't copy): Its particular protagonist, amber portal composition, product branding, UI, level labels, and exact spoken metaphor.
• The transferable format: A personal second-person question, world-building before UI, proof through real app states, progress accumulation, and a human rather than purely functional payoff.
• Use your own cast, location, product capture, visual system, voice, and licensed music.

_{Source: https://x.com/Nlacombe_/status/2106829867463627039 · measured with dissect 1.1.0 on 2026-10-07 · visual estimates and narrative interpretations are marked or described as inferred}