Back Office · office.temerarii.xyz
One asset, all the way in — composition, the wireframe + storyboard, the output format stack, and the template, all read from the SAME content-index record. The expected output matches what /media surfaces for this post.
post social-W24-Sun-4kind threadweek W24date 2026-06-14campaign thread · Sunpillar performancebeat Sunasset videoduration 52.5sground blackscenes 9

Checklist the per-video bar — engine/sim

98.5/100
plain languagevo coverageno dead airuniquenesscaption fitcompletenesscleanliness
quantitative quality · weights learn from your reviews (engine.sim.memory review social-W24-Sun-4 good|bad)
⚠ 1 flag(s) — not yet ship-ready: copy_generic · see docs/strategy/VIDEO-CHECKLIST.md

Composition comp · template family · expected output

composition SceneReelfamily / template SignalFieldReel
9:16 Reelpending1:1 Squarepending16:9 Widepending9:16 4Kpending1:1 4Kpending16:9 4KpendingGIF (SMS)pending
render pending — silent master not yet on disk
expected output: 0/7 rendered — same matrix the /media preview surfaces for this asset.

Composition layer × scene 9 scenes · 52.5s · comp_id + rendered still + tier + the script

#Layer (comp_id · still · tier)BeatTimecodeMotionLogoAudioVO / on-screen / caption
1s1
matches intent
shared field
signature-3d		open0–5.0sspatial-parallaxicon·white-knockout♪ bed_in
Most AI copy reads like a polite robot filling space.
on-screen: AI copy reads like a robot
expected on screen: black ground · dodeca hero in the shared Signal Field · Mensor leads · node-graph · spatial-parallax · icon·white-knockout logo · caption bottom-left
spec (the prompt): comp_id shared Signal Fieldvisual node-graphshape dodecaground blacktreatment white-knockoutmotion spatial-parallaxpower summoninstrument summon→AI copy reads like a robot
2s2
matches intent
shared field
signature-3d		hook5.0–12.3skinetic-buildicon·wireframe♪ node_lock
Here is how The Big T-M gets a model to write copy that sounds like a person who cares.
on-screen: Make AI write like a person
expected on screen: black ground · dodeca hero in the shared Signal Field · Mensor leads · mark · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id shared Signal Fieldvisual markshape dodecaground blacktreatment wireframemotion kinetic-buildpower laser-lockinstrument laser-trace→Make AI write like a person
3s3
matches intent
NumberedList
template		teach12.3–19.6skinetic-buildicon·wireframe♪ node_lock
Paste in real reviews and support chats, then tell the model to steal the phrasing your buyers already use.
on-screen: Feed it real customer words
expected on screen: black ground · a NumberedList panel over a dimmed Signal Field · Mensor leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id NumberedListvisual node-graphshape dodecaground blacktreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→Feed it real customer wordscurate items, nodes
4s4
matches intent
ComparisonTable
template		proof19.6–25.1sreceipts-counticon·wireframe♪ node_lock
The proof: copy built from your customers' own words out-converts clever lines every time.
on-screen: Their words beat clever lines
expected on screen: black ground · a ComparisonTable panel over a dimmed Signal Field · Mensor leads · receipts · receipts-count · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id ComparisonTablevisual receiptsshape dodecaground blacktreatment wireframemotion receipts-countpower receiptsinstrument spotlight→Their words beat clever linescurate colA, colB, rows, statsLabels
5s5
matches intent
BuildLog
template		diff25.1–31.0scrossfade-8ficon·wireframe♪ node_lock
The old way was a thesaurus, a deadline, and a writer's ego in the way.
on-screen: Old way: a thesaurus and ego
expected on screen: black ground · a BuildLog panel over a dimmed Signal Field · Mensor leads · code · crossfade-8f · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id BuildLogvisual codeshape dodecaground blacktreatment wireframemotion crossfade-8fpower morphinstrument morph→Old way: a thesaurus and egocurate codeLines, lines
6s6
matches intent
ChecklistCard
template		teach31.0–36.5skinetic-buildicon·wireframe♪ node_lock
Ask for ten variations, then keep the one that sounds least like a brochure.
on-screen: Write ten, keep the worst-sounding
expected on screen: black ground · a ChecklistCard panel over a dimmed Signal Field · Mensor leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id ChecklistCardvisual node-graphshape dodecaground blacktreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→Write ten, keep the worst-soundingcurate items, nodes
7s7
matches intent
KpiGrid
template		proof36.5–41.5sreceipts-counticon·wireframe♪ node_lock
The receipt is specificity: a real detail beats a smooth, empty promise.
on-screen: Specific beats smooth
expected on screen: black ground · a KpiGrid panel over a dimmed Signal Field · Mensor leads · receipts · receipts-count · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id KpiGridvisual receiptsshape dodecaground blacktreatment wireframemotion receipts-countpower receiptsinstrument spotlight→Specific beats smoothcurate kpis, statsLabels
8s8
matches intent
AnnotatedDiagram
template		step41.5–47.0skinetic-buildicon·wireframe♪ node_lock
Run two headlines live and let the clicks, not your taste, pick the winner.
on-screen: Test two headlines, let clicks vote
expected on screen: black ground · a AnnotatedDiagram panel over a dimmed Signal Field · Mensor leads · pipeline · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id AnnotatedDiagramvisual pipelineshape dodecaground blacktreatment wireframemotion kinetic-buildpower throwinstrument laser-trace→Test two headlines, let clicks votecurate callouts, stages, steps
9s9
matches intent
shared field
signature-3d		resolve47.0–52.5scoalescenceicon·white-knockout♪ bed_out
Let the model draft fast, then you edit the human back into the room.
on-screen: Edit the human back in
expected on screen: black ground · dodeca hero in the shared Signal Field · Mensor leads · coalescence · coalescence · icon·white-knockout logo · caption bottom-left
spec (the prompt): comp_id shared Signal Fieldvisual coalescenceshape dodecaground blacktreatment white-knockoutmotion coalescencepower coalescenceinstrument coalescence→Edit the human back in

Format stack 3 aspects · same scenes[], re-cropped

9:16
1080×1920
Stories · TikTok · YouTube Shorts · Reels
1:1
1080×1080
LinkedIn · Facebook · Instagram
16:9
1920×1080
X/Twitter · YouTube · LinkedIn video

Channels 9 destinations

LinkedInX/TwitterYouTubeInstagramFacebookThreadsTikTokPinterestBluesky

Social captions supplemental published copy · per channel (comp_id level)

tiktokAI copy reads like a polite robot. AI-assisted fix: paste in real reviews and support chats, tell it to steal your buyers' exact phrasing, ask for 10 versions, keep the least brochure-y one, then A/B two headlines. Method below.
instagramAI copy sounds robotic. Fix it: Feed it real reviews. Steal customer phrasing. Write 10, keep the human one. Let clicks vote. #copywriting #marketing #aiwriting #conversion #temerarii
linkedinMost AI copy reads like a polite robot because it's trained on the average, not on your customers. The method we use: paste real reviews and support transcripts into the model and instruct it to reuse the exact phrasing buyers already use, generate ten variations, keep the one that sounds least like a brochure, then A/B two headlines and let clicks decide. Takeaway: copy built from customer language out-converts clever lines.
xAI copy reads like a robot. Fix: feed it real reviews → tell it to steal buyer phrasing → write 10 → keep the least brochure-y → A/B two headlines. Edit the human back in. temerarii.com
facebookMost AI copy reads like a polite robot filling space. Here's the move: paste in real reviews and support chats, tell the model to reuse the exact words your buyers use, ask for ten versions, and keep the one that sounds least like a brochure. Then test two headlines and let the clicks vote. Want the full method? It's on the site.
threadsAI copy reads like a polite robot. Feed it real reviews, tell it to steal buyer phrasing, write 10, keep the least brochure-y one, then A/B two headlines. Edit the human back in.
pinterestAI copywriting that converts: how to write marketing copy using real customer reviews, voice-of-customer phrasing, ten-variation drafting, and headline A/B tests. Plain-language conversion copywriting guide for marketers.
blueskyAI copy reads like a robot. Feed it real reviews, steal buyer phrasing, write 10, keep the human one, A/B two headlines. temerarii.com
youtubeHow To Make AI Write Copy That Sounds Human Most AI copy reads like a polite robot because it's trained on the average. This breakdown shows the method: paste real reviews and support chats into the model, instruct it to reuse your buyers' exact phrasing, generate ten variations, keep the least brochure-like one, and A/B two headlines so clicks pick the winner. AI-assisted production. Give-it-away method from The Big T-M. Chapters: the robot problem · feed it real words · write ten · specificity wins · let clicks vote.

Cross-links every lens is a view on this one record

→ Wireframe · scenes→ Calendar week→ Media (audience preview)→ Storyboard hub→ Day view→ Channel view