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-W24-Tuekind longformweek W24date 2026-06-16campaign longform-youtubepillar multimediabeat asset videoduration 171.3sground whitescenes 7

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-W24-Tue 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 7 scenes · 171.3s · comp_id + rendered still + tier + the script

#Layer (comp_id · still · tier)BeatTimecodeMotionLogoAudioVO / on-screen / caption
1s1
matches intent
shared field
signature-3d
open0–25.9sspatial-parallaxicon·wireframe♪ bed_in
The week is one machine, eight pillars. Today's pillar piece is creative conception, the part everyone thinks a machine cannot do: coming up with the idea. We disagree, with a catch. A model will not hand you a great idea from thin air, but it will generate twenty rough ones fast so you can pick the one worth making. Here is how to run idea generation without it turning to mush.
on-screen: Creative ideas, on demand
expected on screen: white ground · octa hero in the shared Signal Field · Lumen leads · node-graph · spatial-parallax · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id shared Signal Fieldvisual node-graphshape octaground whitetreatment wireframemotion spatial-parallaxpower summoninstrument summon→Creative ideas, on demand
2s2
matches intent
NumberedList
template
teach25.9–50.7skinetic-buildicon·wireframe♪ node_lock
Step one. Before you ask for any idea, paste in your limits. The audience, the one promise, the banned words, the three colors. Ideas get good when they are boxed in, not when they are free. So you tell Claude Code the walls of the room first, then ask it to fill the room. A constrained ask gives you ten usable concepts instead of a hundred useless ones.
on-screen: Feed it the constraints first
expected on screen: white ground · a NumberedList panel over a dimmed Signal Field · Lumen leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id NumberedListvisual node-graphshape octaground whitetreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→Feed it the constraints firstcurate items, nodes
3s3
matches intent
ChecklistCard
template
teach50.7–75.9skinetic-buildicon·wireframe♪ node_lock
Step two. Never ask for the idea. Ask for ten, in a list, one line each. Quantity is the trick, because the first three will be obvious, the middle four boring, and the last three strange enough to be interesting. You are mining for the odd one. Tell the model plainly: give me ten directions, no explanations, and let me choose. You stay the editor; it stays the intern.
on-screen: Ask for ten, not one
expected on screen: white ground · a ChecklistCard panel over a dimmed Signal Field · Lumen leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id ChecklistCardvisual node-graphshape octaground whitetreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→Ask for ten, not onecurate items, nodes
4s4
matches intent
LogStream
template
teach75.9–100.7skinetic-buildicon·wireframe♪ node_lock
Step three. Take your favorite of the ten and ask the model to attack it. Tell it to list why this idea is generic, where it has been done, and what is missing. This is the move people skip. A second pass that tries to break the idea will surface the weak spot while it is still cheap to fix, before you have spent a day building it.
on-screen: Make it argue against itself
expected on screen: white ground · a LogStream panel over a dimmed Signal Field · Lumen leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id LogStreamvisual node-graphshape octaground whitetreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→Make it argue against itselfcurate nodes, rows
5s5
matches intent
WireframeMock
template
teach100.7–125.2skinetic-buildicon·wireframe♪ node_lock
Step four. Once one idea survives, ask the model to write it up as a short brief: the hook, the point, the one thing the viewer should remember. Now your idea is a document anyone, or any other agent, can build from. The conception step ends not with a vague spark but with a written brief that the rest of the studio can pick up and run.
on-screen: Turn the winner into a brief
expected on screen: white ground · a WireframeMock panel over a dimmed Signal Field · Lumen leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id WireframeMockvisual node-graphshape octaground whitetreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→Turn the winner into a briefcurate nodes
6s6
matches intent
KpiGrid
template
proof125.2–147.5sreceipts-counticon·wireframe♪ node_lock
Our proof is the calendar itself. Every week and every day you see was conceived this way, constraints in, a list out, the weak ones cut, the survivor written into a brief. The characters, the angles, the hero lines, none of it was waiting in a drawer. It was generated against the rules and then chosen. The machine proposed; we decided.
on-screen: Every cell on our calendar started here
expected on screen: white ground · a KpiGrid panel over a dimmed Signal Field · Lumen leads · receipts · receipts-count · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id KpiGridvisual receiptsshape octaground whitetreatment wireframemotion receipts-countpower receiptsinstrument spotlight→Every cell on our calendar started herecurate kpis, statsLabels
7s7
matches intent
shared field
signature-3d
resolve147.5–171.3scoalescenceicon·wireframe♪ bed_out
So creative conception with a model is not a mystery, it is volume plus a hard edit. Feed the constraints, ask for ten, make it argue, write the brief. Your next step is easy: open the terminal, paste your one promise, and ask for ten directions, then keep one. You can see what survived our own edit, every day, at office dot temerarii dot xyz.
on-screen: Watch ideas become assets
expected on screen: white ground · octa hero in the shared Signal Field · Lumen leads · coalescence · coalescence · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id shared Signal Fieldvisual coalescenceshape octaground whitetreatment wireframemotion coalescencepower coalescenceinstrument coalescence→Watch ideas become assets

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)

youtubeGet Creative Ideas on Demand by Feeding the Model Constraints First Waiting for inspiration is not a plan. This video shows how to pull usable creative ideas on demand by treating the model as a partner that argues, not an oracle that guesses. What you will learn: - Feed it the constraints first: the audience, the brand rules, the one thing the idea has to do. A model with limits gives sharper ideas than a blank prompt. - Ask for ten, not one: range beats a single guess. You want a spread to choose from, not the first thing it says. - Make it argue against itself: have the model attack its own ideas and name why each would fail, so the weak ones fall out before you spend time on them. - Turn the winner into a brief: take the idea that survives and expand it into a concrete spec the rest of the pipeline can build. Then the proof: every cell on our calendar started here, as an idea pulled this way and pressure-tested against the brand file before it ever became a script. Keywords: idea generation, creative brief, prompt constraints, AI brainstorming, content planning. See the whole system run in public at temerarii.xyz.

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