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-W45-Tuekind longformweek W45date 2026-11-10campaign longform-youtubepillar it_devbeat asset videoduration 128.8sground 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-W45-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 · 128.8s · comp_id + rendered still + tier + the script

#Layer (comp_id · still · tier)BeatTimecodeMotionLogoAudioVO / on-screen / caption
1s1
matches intent
shared field
signature-3d
open0–16.6sspatial-parallaxicon·ember-fill♪ bed_in
A plan is nothing until something gets made. Today is content creation. The social engine needs posts, and lots of them, without burning your whole week. By the end you will know how to make the actual posts with a machine, in your voice, fast.
on-screen: Make the posts, not just plan them
expected on screen: white ground · box hero in the shared Signal Field · Faber leads · node-graph · spatial-parallax · icon·ember-fill logo · caption bottom-left
spec (the prompt): comp_id shared Signal Fieldvisual node-graphshape boxground whitetreatment ember-fillmotion spatial-parallaxpower summoninstrument summon→Make the posts, not just plan them
2s2
matches intent
NumberedList
template
teach16.6–37.5skinetic-buildicon·wireframe♪ node_lock
First move. Before you ask for a single post, give the model a folder of your best past writing. We keep a small corpus of real captions and scripts. The agent reads it and learns your rhythm, short and dry. Now when it drafts, it sounds like you on a calm day, not like every other brand.
on-screen: Feed it how you really talk
expected on screen: white ground · a NumberedList panel over a dimmed Signal Field · Faber leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id NumberedListvisual node-graphshape boxground whitetreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→Feed it how you really talkcurate items, nodes
3s3
matches intent
ChecklistCard
template
teach37.5–57.0skinetic-buildicon·wireframe♪ node_lock
Second move. Tell the model the day and the one idea, and let it draft the post. Because it already holds the source file, it knows Tuesday is a content post on white ground. You are not starting from a blank page. You are editing a first draft that already fits the week.
on-screen: Draft from the day's angle
expected on screen: white ground · a ChecklistCard panel over a dimmed Signal Field · Faber leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id ChecklistCardvisual node-graphshape boxground whitetreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→Draft from the day's anglecurate items, nodes
4s4
matches intent
WireframeMock
template
teach57.0–76.8skinetic-buildicon·wireframe♪ node_lock
Third move. A post needs an image. Wire an image model through MCP, we use a fast generation server, and have the agent write the prompt from the post itself. The words and the picture come from the same idea, so they match. You approve or you nudge one line. No separate design ticket.
on-screen: Make the picture too
expected on screen: white ground · a WireframeMock panel over a dimmed Signal Field · Faber leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id WireframeMockvisual node-graphshape boxground whitetreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→Make the picture toocurate nodes
5s5
matches intent
BlueprintGrid
template
teach76.8–93.8skinetic-buildicon·wireframe♪ node_lock
Fourth move. One draft is many posts. Ask the model to cut the post into a short caption, a longer one, and a three-part thread. Same point, three shapes for three platforms. You wrote once. The machine handled the reshaping that used to eat an afternoon.
on-screen: Cut it into pieces
expected on screen: white ground · a BlueprintGrid panel over a dimmed Signal Field · Faber leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id BlueprintGridvisual node-graphshape boxground whitetreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→Cut it into piecescurate nodes, rows
6s6
matches intent
RankList
template
proof93.8–111.1sreceipts-counticon·wireframe♪ node_lock
Honest proof. Every post in our own week, including this video, was drafted by the same setup. A corpus of our voice, the source file for the angle, an image model for the picture. No outside studio. You are watching the content machine make its own content.
on-screen: This week was made this way
expected on screen: white ground · a RankList panel over a dimmed Signal Field · Faber leads · receipts · receipts-count · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id RankListvisual receiptsshape boxground whitetreatment wireframemotion receipts-countpower receiptsinstrument spotlight→This week was made this waycurate rows, statsLabels
7s7
matches intent
shared field
signature-3d
resolve111.1–128.79999999999998scoalescenceicon·ember-fill♪ bed_out
That is content creation with a machine. Teach it your voice, draft from the plan, make the image, cut the pieces. The work shrinks to taste and approval. Go see the full engine in public at office dot temerarii dot xyz, then point it at your own week.
on-screen: Build your own at the office
expected on screen: white ground · box hero in the shared Signal Field · Faber leads · coalescence · coalescence · icon·ember-fill logo · caption bottom-left
spec (the prompt): comp_id shared Signal Fieldvisual coalescenceshape boxground whitetreatment ember-fillmotion coalescencepower coalescenceinstrument coalescence→Build your own at the office

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)

youtubeMake Social Posts With AI in Your Own Voice: Corpus, Draft, Image, Reshape A plan is nothing until something gets made. In this video we show how to make the actual posts with a machine, in your voice, fast, without burning your whole week. The moves: - Feed it how you really talk. Before you ask for a single post, give the model a folder of your best past writing. We keep a small corpus of real captions and scripts. The agent reads it and learns your rhythm, short and dry. - Draft from the day's angle. Tell the model the day and the one idea and let it draft the post. Because it already holds the source file, it knows the day's ground and theme. You edit a first draft that already fits the week. - Make the picture too. Wire an image model through MCP and have the agent write the prompt from the post itself. The words and the picture come from the same idea, so they match. No separate design ticket. - Cut it into pieces. Ask the model to cut the post into a short caption, a longer one, and a three-part thread. Same point, three shapes for three platforms. The honest proof: every post in our own week, including this video, was drafted by the same setup. A corpus of our voice, the source file for the angle, an image model for the picture. No outside studio. You are watching the content machine make its own content. The takeaway: teach it your voice, draft from the plan, make the image, cut the pieces. The work shrinks to taste and approval. See the full engine at office.temerarii.xyz. Keywords: AI content creation, brand voice, image generation, MCP, social media posts.

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