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-W44-Monkind longformweek W44date 2026-11-02campaign longform-youtubepillar socialbeat asset videoduration 173.6sground 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-W44-Mon 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 · 173.6s · comp_id + rendered still + tier + the script

#Layer (comp_id · still · tier)BeatTimecodeMotionLogoAudioVO / on-screen / caption
1s1
matches intent
shared field
signature-3d
open0–23.0sspatial-parallaxicon·ember-fill♪ bed_in
Today we go deep on one thing: social listening. The calendar this week is about building presence in public, and you cannot build for a room you have never heard. Most accounts shout into the dark. By the end of this you will know how to pull what your audience actually says, with a real tool, and turn it into posts that land.
on-screen: Listen before you post
expected on screen: black ground · torus hero in the shared Signal Field · Nuntius leads · node-graph · spatial-parallax · icon·ember-fill logo · caption bottom-left
spec (the prompt): comp_id shared Signal Fieldvisual node-graphshape torusground blacktreatment ember-fillmotion spatial-parallaxpower summoninstrument summon→Listen before you post
2s2
matches intent
NumberedList
template
teach23.0–44.6skinetic-buildicon·ember-fill♪ node_lock
First step. Pick a place your people gather. For a lot of work, that is Reddit. We wire a Reddit MCP server into the agent, which is just an adapter that lets the model read posts and comments in a subreddit directly. No scraping by hand, no copy-paste. The model can now sit in the room and take notes.
on-screen: Wire an MCP into Reddit
expected on screen: black ground · a NumberedList panel over a dimmed Signal Field · Nuntius leads · node-graph · kinetic-build · icon·ember-fill logo · caption bottom-left
spec (the prompt): comp_id NumberedListvisual node-graphshape torusground blacktreatment ember-fillmotion kinetic-buildpower morphinstrument morph+laser→Wire an MCP into Redditcurate items, nodes
3s3
matches intent
ChecklistCard
template
teach44.6–65.9skinetic-buildicon·wireframe♪ node_lock
Second step. We do not ask for likes or follower counts. We ask the model to pull the questions people keep asking and the complaints they keep repeating. Those repeats are the signal. A question asked once is noise. A question asked forty times is a post you have to write. The model sorts the repeats for you.
on-screen: Pull the real questions, not vanity
expected on screen: black ground · a ChecklistCard panel over a dimmed Signal Field · Nuntius leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id ChecklistCardvisual node-graphshape torusground blacktreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→Pull the real questions, not vanitycurate items, nodes
4s4
matches intent
AnnotatedDiagram
template
teach65.9–88.60000000000001skinetic-buildicon·white-knockout♪ node_lock
Third step. One forum is too small a window. So we add a web search tool through the same agent and ask it what is being said about our topic this week across the open web. Now the model is reading two rooms at once, the forum and the wider feed, and it tells you where they agree and where they fight.
on-screen: Add web search for the wider room
expected on screen: black ground · a AnnotatedDiagram panel over a dimmed Signal Field · Nuntius leads · node-graph · kinetic-build · icon·white-knockout logo · caption bottom-left
spec (the prompt): comp_id AnnotatedDiagramvisual node-graphshape torusground blacktreatment white-knockoutmotion kinetic-buildpower morphinstrument morph+laser→Add web search for the wider roomcurate callouts, nodes
5s5
matches intent
WireframeMock
template
teach88.6–110.89999999999999skinetic-buildicon·liquid-chrome♪ node_lock
Fourth step. Raw chatter is a mess. We ask the model to group everything it found into a handful of plain themes, and to name each theme in one short line a regular person would say out loud. Now you are not staring at a thousand comments. You are staring at five themes, each one a post waiting to be written.
on-screen: Cluster the noise into themes
expected on screen: black ground · a WireframeMock panel over a dimmed Signal Field · Nuntius leads · node-graph · kinetic-build · icon·liquid-chrome logo · caption bottom-left
spec (the prompt): comp_id WireframeMockvisual node-graphshape torusground blacktreatment liquid-chromemotion kinetic-buildpower morphinstrument morph+laser→Cluster the noise into themescurate nodes
6s6
matches intent
BuildLog
template
teach110.9–131.1skinetic-buildicon·ember-fill♪ node_lock
Fifth step. We take those five themes and drop them straight back into the source file as next week's topics. The thing the audience said becomes the thing we make. The loop closes. Listening is not a side task you do when you have time. It is the front of the line, feeding everything downstream.
on-screen: Themes become next week's outline
expected on screen: black ground · a BuildLog panel over a dimmed Signal Field · Nuntius leads · node-graph · kinetic-build · icon·ember-fill logo · caption bottom-left
spec (the prompt): comp_id BuildLogvisual node-graphshape torusground blacktreatment ember-fillmotion kinetic-buildpower morphinstrument morph+laser→Themes become next week's outlinecurate lines, nodes
7s7
matches intent
StatScoreboard
template
proof131.1–153.1sreceipts-counticon·wireframe♪ node_lock
The honest proof. We run this exact loop on our own work, including the room around The Big T-M, and the questions people ask us are what shaped this very week. We are not guessing what you want to hear. We pulled it. You can see the listening output and the calendar it built at office dot temerarii dot xyz.
on-screen: We listen to our own room too
expected on screen: black ground · a StatScoreboard panel over a dimmed Signal Field · Nuntius leads · receipts · receipts-count · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id StatScoreboardvisual receiptsshape torusground blacktreatment wireframemotion receipts-countpower receiptsinstrument spotlight→We listen to our own room toocurate pillar, stats, statsLabels
8s8
matches intent
shared field
signature-3d
resolve153.1–173.6scoalescenceicon·ember-fill♪ bed_out
The takeaway is simple. Stop posting at people and start answering them. Wire a listening tool into your agent, pull the repeated questions, name the themes, and write to those. The room is already telling you what to make. Go set up your first listening pull, and watch ours run at office dot temerarii dot xyz.
on-screen: Hear the room, then answer it
expected on screen: black ground · torus hero in the shared Signal Field · Nuntius leads · coalescence · coalescence · icon·ember-fill logo · caption bottom-left
spec (the prompt): comp_id shared Signal Fieldvisual coalescenceshape torusground blacktreatment ember-fillmotion coalescencepower coalescenceinstrument coalescence→Hear the room, then answer it

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)

youtubeHow to carry one story across every platform without copy pasting Posting the same clip everywhere is lazy, and re writing it five times by hand is a waste. This video shows how to carry one story across every platform in each platform's native shape, with a system doing the reshaping. The method: Write the story once as the source. One real point, in your own voice, with the method given away. That source is the truth every platform version pulls from, so they all say the same thing even when they look different. Reshape per platform from that source, do not retype it. A model takes the one story and casts it native to each place, the aspect, the length, the caption style, while the point stays the same. A vertical clip and a long post are the same story wearing different clothes. Pin the rules each platform needs. Aspect ratio, caption length, where the link goes, the thumbnail. Keep those as data the system honors, so every version comes out correct without you remembering each platform's quirks. Tag every version at publish so you can tell which platform carried the story. The tag names the campaign and the channel, so the result traces back to the right place instead of a guess. Keep the voice one voice. The same brand rules and banned words run on every platform version, so the story sounds like you whether it is fifteen seconds or fifteen hundred words. One story, many native shapes, one voice, traced back to the source. That is how a small team shows up everywhere without copy pasting itself into mush. See one story hold across every channel at office.temerarii.xyz. Keywords: content repurposing, cross platform, social media, native content, omnichannel, distribution.

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