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-W30-Wedkind longformweek W30date 2026-07-29campaign longform-youtubepillar socialbeat asset videoduration 188.9sground blackscenes 8

Checklist the per-video bar — engine/sim

98.4/100
plain languagevo coverageno dead airuniquenesscaption fitcompletenesscleanliness
quantitative quality · weights learn from your reviews (engine.sim.memory review longform-W30-Wed good|bad)
⚠ 1 flag(s) — not yet ship-ready: copy_generic · see docs/strategy/VIDEO-CHECKLIST.md

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 · 188.9s · comp_id + rendered still + tier + the script

#Layer (comp_id · still · tier)BeatTimecodeMotionLogoAudioVO / on-screen / caption
1s1
matches intent
shared field
signature-3d
open0–29.8sspatial-parallaxicon·ember-fill♪ bed_in
One idea, nine outputs, and today the scary one: live streaming. Live has no second take. That is exactly why it builds trust faster than a polished clip, because nobody can fake it in real time. We are going to show you how to run a live show that does not fall apart, with a model helping behind the glass. By the end you will know how to plan a stream that feels loose but is actually built on a tight skeleton.
on-screen: Live with nothing hidden
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→Live with nothing hidden
2s2
matches intent
NumberedList
template
teach29.8–57.5skinetic-buildicon·wireframe♪ node_lock
Live does not want a word-for-word script, it wants a rundown. A rundown is a list of segments with a time on each: two minutes hello, eight minutes the demo, five minutes questions. We keep it in the same file as the rest of the week, so the live show says the same one idea as the video and the photos. The rundown is your map when your brain goes blank on camera, and it will.
on-screen: Write a rundown, not a script
expected on screen: black ground · a NumberedList panel over a dimmed Signal Field · Nuntius leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id NumberedListvisual node-graphshape torusground blacktreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→Write a rundown, not a scriptcurate items, nodes
3s3
matches intent
ChecklistCard
template
teach57.5–80.2skinetic-buildicon·wireframe♪ node_lock
You cannot read chat and talk at the same time. So put a model on the chat. It reads every message, groups the repeats, and surfaces the three real questions instead of the noise. You glance at a clean list, not a waterfall. The machine handles volume, you handle the answer. That is the whole division of labor: it filters, you speak.
on-screen: Let a model watch the chat
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→Let a model watch the chatcurate items, nodes
4s4
matches intent
BlueprintGrid
template
teach80.2–103.6skinetic-buildicon·wireframe♪ node_lock
Here is how one stream becomes many outputs. The recording goes to a model with the timestamps from your rundown. It finds the strong moments and marks the start and end of each clip. You confirm, then a tool trims them. By the next morning the live hour has become five short videos. You streamed once and fed the whole week without filming again.
on-screen: Cut the stream into clips, same day
expected on screen: black ground · a BlueprintGrid panel over a dimmed Signal Field · Nuntius leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id BlueprintGridvisual node-graphshape torusground blacktreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→Cut the stream into clips, same daycurate nodes, rows
5s5
matches intent
ListCard
template
teach103.6–128.4skinetic-buildicon·wireframe♪ node_lock
Live breaks. The audio drops, the screen freezes, your mind goes empty. So write the recovery into the rundown before you go on. One filler segment you can always run. One line you say while you fix things. We even keep a slide ready that says, give us a second. Calm on camera is not luck, it is having already decided what to do when it goes wrong.
on-screen: Have a plan for when it breaks
expected on screen: black ground · a ListCard panel over a dimmed Signal Field · Nuntius leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id ListCardvisual node-graphshape torusground blacktreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→Have a plan for when it breakscurate items, nodes
6s6
matches intent
LogStream
template
teach128.4–151.4skinetic-buildicon·wireframe♪ node_lock
We run our own studio out loud, mistakes and all, on a rundown tied to the same source file as this video. The clips you see midweek are cut from those streams by a model reading our own timestamps. Nothing staged, nothing hidden. That is the proof. If we trusted live enough to run our whole operation on it, the method is real.
on-screen: We stream the studio in public
expected on screen: black ground · a LogStream panel over a dimmed Signal Field · Nuntius leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id LogStreamvisual node-graphshape torusground blacktreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→We stream the studio in publiccurate nodes, rows
7s7
matches intent
StatScoreboard
template
proof151.4–175.5sreceipts-counticon·wireframe♪ node_lock
So the takeaway. Live earns trust because it cannot be faked, but only if you build the skeleton first. Write a rundown, put a model on the chat, cut the recording into clips the same day, and decide your recovery before you start. You can see the stream archive and the clips it fed at office dot temerarii dot xyz. Write the rundown, then go live.
on-screen: See the stream archive live
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→See the stream archive livecurate pillar, stats, statsLabels
8s8
matches intent
shared field
signature-3d
resolve175.5–188.9scoalescenceicon·ember-fill♪ bed_out
Chapter seven. Social at scale: one story becomes nine native channels. Write the master script once, then let a render pipeline slice each platform cut and a posting tool drop each one in its home feed.
on-screen: Chapter 7
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→Chapter 7

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)

youtubeRun a Live Stream That Builds Trust: Rundown, a Model on Chat, Same-Day Clips, a Recovery Plan Live has no second take. That is exactly why it builds trust faster than a polished clip, because nobody can fake it in real time. This walkthrough shows how to run a live stream that earns trust and feeds the rest of your content. The method: - Write a rundown, not a script. Live wants a list of segments with a time on each, two minutes hello, eight minutes the demo, five minutes questions. Keep it in the same plan file as the rest of the week. - Let a model watch the chat. You cannot read chat and talk at once. Put a model on the chat, it reads every message, groups the repeats, and surfaces the three real questions instead of the noise. You glance at a clean list, not a waterfall. - Cut the stream into clips, same day. Send the recording to a model with the timestamps from your rundown. It finds the strong moments and marks the start and end of each clip. You confirm, then a tool trims them, and by next morning one stream has become many outputs. - Have a plan for when it breaks. Live breaks, the audio drops, the screen freezes, your mind goes empty. Write the recovery into the rundown before you go on, one filler segment you can always run, one line to say while you fix things. We run our own studio out loud, mistakes and all, on a rundown tied to the same source file as this video. The midweek clips are cut from those streams by a model reading our own timestamps. Nothing staged, nothing hidden. Live earns trust because it cannot be faked, but only if you build the skeleton first. See the stream archive at office.temerarii.xyz. Keywords: live streaming, rundown, chat moderation, clip cutting, content repurposing, video production.

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