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 longform-W31-Frikind longformweek W31date 2026-08-07campaign longform-youtubepillar performancebeat asset videoduration 143.6sground blackscenes 8

Checklist the per-video bar — engine/sim

100.0/100
plain languagevo coverageno dead airuniquenesscaption fitcompletenesscleanliness
quantitative quality · weights learn from your reviews (engine.sim.memory review longform-W31-Fri good|bad)
✓ all static checks pass — one-focal/scene · tier-by-beat · one-track caption · colorway · cast+shape correct · no banned/fabricated. (audio + visual tiers verify on the rendered finals — Phase 2)

Composition comp · template family · expected output

composition LongFormChaptersfamily / template LongFormChapters
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 8 scenes · 143.6s · comp_id + rendered still + tier + the script

#Layer (comp_id · still · tier)BeatTimecodeMotionLogoAudioVO / on-screen / caption
1s1
matches intent
shared field
signature-3d
open0–20.2sspatial-parallaxicon·wireframe♪ bed_in
Cut busywork, keep revenue. Today it is copywriting, where most effort goes into clever lines nobody asked for. We are going to show you how to use a model to write copy grounded in your customers' own words, then test it instead of trusting it. By the end you will write less and convert more.
on-screen: Words that move money
expected on screen: black ground · dodeca hero in the shared Signal Field · Mensor leads · node-graph · spatial-parallax · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id shared Signal Fieldvisual node-graphshape dodecaground blacktreatment wireframemotion spatial-parallaxpower summoninstrument summon→Words that move money
2s2
matches intent
NumberedList
template
teach20.2–37.5skinetic-buildicon·wireframe♪ node_lock
Step one. Before writing a word, collect what real customers say, support tickets, reviews, sales call notes. Feed that pile to Claude Code and ask it to pull the exact phrases people use to describe their problem. Now your copy starts from their language, not your imagination.
on-screen: Mine the words customers use
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→Mine the words customers usecurate items, nodes
3s3
matches intent
ChecklistCard
template
teach37.5–54.8skinetic-buildicon·wireframe♪ node_lock
Step two. Tell the model to write the headline about the reader's problem, not your product. People scan for themselves first. The model is happy to drop the clever wordplay and say the plain thing. Plain and about them beats clever and about you, every single time.
on-screen: Lead with the problem, not you
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→Lead with the problem, not youcurate items, nodes
4s4
matches intent
BlueprintGrid
template
teach54.8–73.6skinetic-buildicon·wireframe♪ node_lock
Step three. Hand your draft back to the model and tell it to remove any sentence that does not help the reader decide. Copy gets fat with throat-clearing. The model trims without ego. What is left is shorter, clearer, and easier to act on, which is the whole job of copy.
on-screen: Cut every word that floats
expected on screen: black ground · a BlueprintGrid panel over a dimmed Signal Field · Mensor leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id BlueprintGridvisual node-graphshape dodecaground blacktreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→Cut every word that floatscurate nodes, rows
5s5
matches intent
StepFlow
template
teach73.6–91.6skinetic-buildicon·wireframe♪ node_lock
Step four. Have the model write two honest versions of the same page with different angles. Put each in front of a slice of traffic and keep the one that converts. The model makes variants cheap, so you stop arguing in a meeting and let real readers settle it.
on-screen: Write two, let readers pick
expected on screen: black ground · a StepFlow panel over a dimmed Signal Field · Mensor leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id StepFlowvisual node-graphshape dodecaground blacktreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→Write two, let readers pickcurate nodes, steps
6s6
matches intent
WireframeMock
template
teach91.6–110.69999999999999skinetic-buildicon·wireframe♪ node_lock
Step five. Ask the model to rewrite the final copy so a child could follow it. Simple is not dumb, it is respect for a busy reader. If a kid gets it, your customer gets it faster. The model is good at stripping jargon down to something a human can actually use.
on-screen: Read it to an eleven-year-old
expected on screen: black ground · a WireframeMock panel over a dimmed Signal Field · Mensor leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id WireframeMockvisual node-graphshape dodecaground blacktreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→Read it to an eleven-year-oldcurate nodes
7s7
matches intent
StatScoreboard
template
proof110.7–127.30000000000001sreceipts-counticon·wireframe♪ node_lock
Every line you read from us comes out of this loop, customer words in, a model drafting, then a human cutting hard. We do not hide a copy department. We hold our own pages to the same plain bar we are handing you right now.
on-screen: We write our own pages this way
expected on screen: black ground · a StatScoreboard panel over a dimmed Signal Field · Mensor leads · receipts · receipts-count · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id StatScoreboardvisual receiptsshape dodecaground blacktreatment wireframemotion receipts-countpower receiptsinstrument spotlight→We write our own pages this waycurate pillar, stats, statsLabels
8s8
matches intent
shared field
signature-3d
resolve127.3–143.6scoalescenceicon·wireframe♪ bed_out
So the takeaway. Good copy is built from your customers' words, kept plain, and tested instead of trusted, with a model speeding up the boring parts. That is how words keep revenue. Go see ours running in public at office dot temerarii dot xyz.
on-screen: See the copy live
expected on screen: black ground · dodeca hero in the shared Signal Field · Mensor leads · coalescence · coalescence · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id shared Signal Fieldvisual coalescenceshape dodecaground blacktreatment wireframemotion coalescencepower coalescenceinstrument coalescence→See the copy live

Format stack 1 aspects · same scenes[], re-cropped

16:9
1920×1080
X/Twitter · YouTube · LinkedIn video

Channels 2 destinations

YouTubeBlog

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

youtubeCopywriting From Your Customers' Own Words: Mine, Lead With the Problem, Cut, and Test Most copy effort goes into clever lines nobody asked for. This walkthrough shows how to use a model to write copy grounded in your customers' own words, then test it instead of trusting it. Words that move money, not your ego. The method: - Mine the words customers use. Before writing, collect what real customers say, support tickets, reviews, sales call notes. Feed that pile to a coding agent and ask it to pull the exact phrases people use to describe their problem. Now your copy starts from their language, not yours. - Lead with the problem, not you. Tell the model to write the headline about the reader's problem, not your product. People scan for themselves first. Plain and about them beats clever and about you. - Cut every word that floats. Hand your draft back and tell the model to remove any sentence that does not help the reader decide. Copy gets fat with throat-clearing. The model trims without ego, and what is left is shorter and easier to act on. - Write two, let readers pick. Have the model write two honest versions with different angles. Put each in front of a slice of traffic and keep the one that converts. The model makes variants cheap, so you stop arguing in a meeting and let readers decide. - Read it to an eleven-year-old. Ask the model to rewrite the final copy so a child could follow it. Simple is not dumb, it is respect for a busy reader. Every line you read from us comes out of this loop, customer words in, a model drafting, then a human cutting hard. We hold our own pages to the same plain bar. Good copy is built from your customers' words, kept plain, and tested instead of trusted. See ours running at office.temerarii.xyz. Keywords: copywriting, voice of customer, landing page copy, A/B testing, plain language, conversion copy.

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