Persistent-Chat Before-After Film
Avant les ré-explications, après le fil qui se souvient.
Transforme votre IA en réalisateur de vidéos de lancement pour un produit de chat persistant — elle scénarise la transformation avant/après, planifie les plans et vous remet un plan de montage prêt à tourner pour un film de lancement de 60 à 90 secondes.

Installez-la en 3 étapes
Sans inscription. Sans outils en ligne de commande. Choisissez votre IA, copiez une ligne, collez-la.
Vous préférez le faire vous-même ? Achetez la compétence ($4.99), téléchargez le ZIP depuis votre confirmation d'achat, décompressez-le dans le répertoire indiqué ci-dessus, redémarrez votre session IA — même résultat. Vous l'avez déjà achetée ? Votre assistant IA peut la télécharger pour vous.
Contenu du pack
Un fichier SKILL.md (le guide pratique) + un modèle de départ. Aperçu :
Persistent-Chat Before-After Film
When to use
- You are launching a chat product where conversations persist across sessions, and the demo needs to show why that matters in under 90 seconds
- Your current launch video shows features one by one, and viewers still ask "so what does it actually do for me?"
- You have real before/after material — messy scattered chats vs. one continuous thread — but no story structure to hold it together
- You need a script plus shot list you can hand to an editor, a motion designer, or a founder filming their own screen
- Your launch is going out on a landing page, a product post, and social clips, and you want one film that cuts down cleanly into all three
Workflow
- Define the transformation (input: product description → output: one-sentence transformation statement)
State the change in one sentence: "Before, every chat started from zero; after, every conversation picks up where it left off." Name the pain in the before-state (re-explaining, lost context, repeated questions) and the payoff in the after-state (continuity, memory, momentum). If you cannot state the transformation in one sentence, the film will ramble — keep rewriting until it fits on a sticky note. Everything downstream — script, shots, captions — must serve this one sentence. - Collect the before/after evidence (input: product access + user stories → output: evidence pack)
Gather concrete material for both sides: screen recordings of the old way (starting over, re-typing context, hunting for a past answer) and the new way (reopening a thread, the chat recalling earlier details, continuing mid-task). Pull two or three real user moments — a returning customer, a multi-day project, a handoff between teammates — and note the exact lines of dialogue or on-screen moments that prove the difference. No evidence, no film: invented before/after moments read as advertising and get skipped. - Script the 60–90 second arc (input: transformation statement + evidence pack → output: timed script)
Structure the film in three beats: Before (0–20s, the frustration, shown not told), Turning point (20–35s, the moment the persistent thread appears), After (35–75s, the payoff across two or three real moments), Close (75–90s, name + one call to action). Write the voiceover or on-screen text to match, timed to the second, with each line tied to a specific shot. Read it aloud twice — if it runs long, cut lines, never speed up the delivery. - Build the shot list (input: timed script → output: shot-by-shot capture plan)
Turn every script line into a shot: screen capture, talking head, product close-up, or text card. For each shot, specify framing, duration, what is on screen, and how it is captured (real screen recording vs. staged mock, cursor visible or hidden, real data or placeholder). Mark the two or three hero shots — the exact frames where the before becomes the after — and plan them first, because they carry the film. Keep the list to 12–18 shots; more than that and the edit turns into a montage nobody follows. - Plan the edit and the cut-downs (input: shot list → output: edit plan + cut-down specs)
Specify pacing (quick cuts in the Before, longer holds in the After), captions (burned-in, word-for-word, readable on mute), music (one track, low under voiceover, silent under the turning point), and the end card (product name, one line, one button). Then define the cut-downs: a 15-second social teaser (turning point + one payoff + end card), a 30-second version (before, turning point, one payoff), and a muted autoplay version for the landing page (captions on, no voiceover dependency). The editor should be able to cut all three from the same footage without re-shooting anything.
Aperçu du modèle
Persistent-Chat Before-After Film — Production Brief
One film, one transformation. Fill this in before scripting anything.
Transformation statement
Before: {what is frustrating about starting over} After: {what is better when the conversation persists} One sentence: "Before, {pain}; after, {payoff}."
Evidence pack
Before moments (the old way)
- {moment 1 — what the user has to redo}
- {moment 2}
After moments (the persistent way)
- {moment 1 — the thread picking up where it left off}
- {moment 2}
Real user moments
Privacy line (one plain sentence): {how history is stored and who can see it}
Timed script (60–90 seconds)
Le pack de compétence lui-même (SKILL.md + modèle) est rédigé en anglais.


