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-Frikind longformweek W30date 2026-07-31campaign longform-youtubepillar staff_trainingbeat asset videoduration 173.8sground whitescenes 7

Checklist the per-video bar — engine/sim

98.2/100
plain languagevo coverageno dead airuniquenesscaption fitcompletenesscleanliness
quantitative quality · weights learn from your reviews (engine.sim.memory review longform-W30-Fri 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 7 scenes · 173.8s · comp_id + rendered still + tier + the script

#Layer (comp_id · still · tier)BeatTimecodeMotionLogoAudioVO / on-screen / caption
1s1
matches intent
shared field
signature-3d
open0–27.7sspatial-parallaxicon·3d-extrude♪ bed_in
One idea, nine outputs, and today the output is a conversation: the interview. A good interview gives you something you cannot script, a real person saying a true thing in their own words. A bad one is a list of yes-or-no answers and dead air. We are going to fix that. By the end you will know how to prep, run, and cut an interview so it carries the same idea as everything else this week.
on-screen: Get someone to talk
expected on screen: white ground · knot hero in the shared Signal Field · Magister leads · node-graph · spatial-parallax · icon·3d-extrude logo · caption bottom-left
spec (the prompt): comp_id shared Signal Fieldvisual node-graphshape knotground whitetreatment 3d-extrudemotion spatial-parallaxpower summoninstrument summon→Get someone to talk
2s2
matches intent
NumberedList
template
teach27.7–51.8skinetic-buildicon·wireframe♪ node_lock
Prep first. Feed the model whatever the guest has already said in public, their posts, their talks, their site. Ask it to find the threads they keep returning to, and the thing they have never been asked. Now you walk in knowing where the gold is. You are not reading a generic question list. You are aiming at the one story only this person can tell.
on-screen: Research the guest with a model
expected on screen: white ground · a NumberedList panel over a dimmed Signal Field · Magister leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id NumberedListvisual node-graphshape knotground whitetreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→Research the guest with a modelcurate items, nodes
3s3
matches intent
ChecklistCard
template
teach51.8–76.3skinetic-buildicon·wireframe♪ node_lock
In the room, the skill is the question and the silence after it. Ask open questions that start with how or why, never ones that can be answered with yes. Then stop talking. The best lines come three seconds into the pause, when the guest fills the quiet. Most interviewers ruin the moment by jumping in. Your job is to ask, then get out of the way.
on-screen: Ask open, then shut up
expected on screen: white ground · a ChecklistCard panel over a dimmed Signal Field · Magister leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id ChecklistCardvisual node-graphshape knotground whitetreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→Ask open, then shut upcurate items, nodes
4s4
matches intent
ListCard
template
teach76.3–102.19999999999999skinetic-buildicon·wireframe♪ node_lock
After the shoot, run the audio through a speech-to-text model so you have every word as text in minutes. Then read the transcript and mark the three lines that made you lean in. Those lines are your spine. You cut the interview to serve them, the same way a script serves its spine. The transcript turns a vague hour of talking into a clear set of moments you can build around.
on-screen: Transcribe, then find the spine
expected on screen: white ground · a ListCard panel over a dimmed Signal Field · Magister leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id ListCardvisual node-graphshape knotground whitetreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→Transcribe, then find the spinecurate items, nodes
5s5
matches intent
SchematicCard
template
teach102.2–123.80000000000001skinetic-buildicon·wireframe♪ node_lock
We use this same loop on ourselves, answering hard questions about how the studio runs on camera, transcribed by a model and cut to the lines that matter. No invented quotes, no scrubbed answers. That is the proof. We do not hide behind a brochure. We sit down and answer, and we cut to the truth, not the flattery.
on-screen: We run our own answers in public
expected on screen: white ground · a SchematicCard panel over a dimmed Signal Field · Magister leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id SchematicCardvisual node-graphshape knotground whitetreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→We run our own answers in publiccurate nodes
6s6
matches intent
KpiGrid
template
proof123.8–149.0sreceipts-counticon·wireframe♪ node_lock
So the takeaway. An interview is research plus silence. Use a model to find the one story only this guest can tell, ask open questions and then wait, transcribe the whole thing, and cut to the three lines that made you lean in. You can watch the interviews from this week at office dot temerarii dot xyz. Do the homework, then ask one good question and let them talk.
on-screen: Prep deep, then ask and wait
expected on screen: white ground · a KpiGrid panel over a dimmed Signal Field · Magister leads · receipts · receipts-count · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id KpiGridvisual receiptsshape knotground whitetreatment wireframemotion receipts-countpower receiptsinstrument spotlight→Prep deep, then ask and waitcurate kpis, statsLabels
7s7
matches intent
shared field
signature-3d
resolve149.0–173.8scoalescenceicon·3d-extrude♪ bed_out
And here is the bonus that ties back to the week. One good interview is not one output, it is many. The same transcript gives you the long conversation, a handful of short clips, and pull quotes for the posts. You sat down once. The machine helped you find the moments, and one honest conversation fed the whole calendar. That is what one idea, nine outputs really means.
on-screen: One conversation, many clips
expected on screen: white ground · knot hero in the shared Signal Field · Magister leads · coalescence · coalescence · icon·3d-extrude logo · caption bottom-left
spec (the prompt): comp_id shared Signal Fieldvisual coalescenceshape knotground whitetreatment 3d-extrudemotion coalescencepower coalescenceinstrument coalescence→One conversation, many clips

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 Run an Interview That Gives You More Than Yes-or-No Answers A good interview gives you something you cannot script, a real person saying a true thing in their own words. A bad one is a list of yes-or-no answers and dead air. This walkthrough shows how to get the good kind, and turn one conversation into many outputs. The method: - Research the guest with a model. Feed it whatever the guest has already said in public, their posts, talks, site. Ask it to find the threads they keep returning to, and the thing they have never been asked. Now you walk in knowing where the gold is. - Ask open, then shut up. In the room, the skill is the question and the silence after it. Ask open questions that start with how or why, never ones answered with yes. Then stop talking. The best lines come three seconds into the pause, when the guest fills the quiet. - Transcribe, then find the spine. Run the audio through a speech-to-text model so you have every word as text in minutes. Read the transcript and mark the three lines that made you lean in. Those lines are your spine, and you cut the interview to serve them. We run this loop on ourselves, answering hard questions about how the studio runs on camera, transcribed and cut to the lines that matter. No invented quotes, no scrubbed answers. The bonus: one good interview is not one output, it is many. The same transcript gives you the long conversation, short clips, and pull quotes for posts. You sat down once. See it at office.temerarii.xyz. Keywords: interview technique, open questions, speech to text, transcript, content repurposing, video production.

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