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-W29-Wedkind longformweek W29date 2026-07-22campaign longform-youtubepillar multimediabeat asset videoduration 119.9sground 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-W29-Wed 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 · 119.9s · comp_id + rendered still + tier + the script

#Layer (comp_id · still · tier)BeatTimecodeMotionLogoAudioVO / on-screen / caption
1s1
matches intent
shared field
signature-3d
open0–18.0sspatial-parallaxicon·wireframe♪ bed_in
You have data. You probably also have ten dashboards nobody reads. Today, business intelligence: how to make the agent read your own numbers and tell you, in a sentence, what changed and what to do. The goal is an answer, not another chart you scroll past on a Monday.
on-screen: Stop staring at dashboards
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→Stop staring at dashboards
2s2
matches intent
NumberedList
template
teach18.0–35.3skinetic-buildicon·color♪ node_lock
First, give the agent the numbers, not a screenshot of the numbers. Wire an MCP server into your analytics, like Google Search Console or your database, and let the model pull the rows directly. Now it is reading the live source, the same one your dashboard reads.
on-screen: Point an agent at your real data
expected on screen: white ground · a NumberedList panel over a dimmed Signal Field · Lumen leads · node-graph · kinetic-build · icon·color logo · caption bottom-left
spec (the prompt): comp_id NumberedListvisual node-graphshape octaground whitetreatment colormotion kinetic-buildpower morphinstrument morph+laser→Point an agent at your real datacurate items, nodes
3s3
matches intent
ChecklistCard
template
teach35.3–53.0skinetic-buildicon·color♪ node_lock
Instead of building a report, ask. Type: which pages lost the most search clicks this month, and why. The agent queries the data, does the math, and answers in plain language. You skipped the part where a human exports to a spreadsheet and color-codes it for an hour.
on-screen: Ask the question in plain words
expected on screen: white ground · a ChecklistCard panel over a dimmed Signal Field · Lumen leads · node-graph · kinetic-build · icon·color logo · caption bottom-left
spec (the prompt): comp_id ChecklistCardvisual node-graphshape octaground whitetreatment colormotion kinetic-buildpower morphinstrument morph+laser→Ask the question in plain wordscurate items, nodes
4s4
matches intent
LegacyCard
template
teach53.0–71.4skinetic-buildicon·color♪ node_lock
Push it one step further. Ask the agent to turn its own finding into an action: rewrite the three weakest page titles using the queries people actually searched. It reads your real query data and drafts the fix. The insight and the next move come out in the same run.
on-screen: Let it write the next step, not
expected on screen: white ground · a LegacyCard panel over a dimmed Signal Field · Lumen leads · node-graph · kinetic-build · icon·color logo · caption bottom-left
spec (the prompt): comp_id LegacyCardvisual node-graphshape octaground whitetreatment colormotion kinetic-buildpower morphinstrument morph+laser→Let it write the next step, not just the chartcurate nodes
5s5
matches intent
StepFlow
template
teach71.4–88.4skinetic-buildicon·color♪ node_lock
Write that question into a workflow file so it runs every Monday on its own. Same query, fresh data, a short answer in your inbox before you open your laptop. You turned a recurring report into a recurring answer, which is the only kind anyone reads.
on-screen: Save the question, run it weekly
expected on screen: white ground · a StepFlow panel over a dimmed Signal Field · Lumen leads · node-graph · kinetic-build · icon·color logo · caption bottom-left
spec (the prompt): comp_id StepFlowvisual node-graphshape octaground whitetreatment colormotion kinetic-buildpower morphinstrument morph+laser→Save the question, run it weeklycurate nodes, steps
6s6
matches intent
ComparisonTable
template
proof88.4–105.4sreceipts-counticon·color♪ node_lock
Proof without a figure: we point this at our own site's search data and let the model tell us which pages to fix next, in public. The pages you will see were edited based on what real searchers typed, not on a hunch in a meeting.
on-screen: We read our own traffic this way
expected on screen: white ground · a ComparisonTable panel over a dimmed Signal Field · Lumen leads · receipts · receipts-count · icon·color logo · caption bottom-left
spec (the prompt): comp_id ComparisonTablevisual receiptsshape octaground whitetreatment colormotion receipts-countpower receiptsinstrument spotlight→We read our own traffic this waycurate colA, colB, rows, statsLabels
7s7
matches intent
shared field
signature-3d
resolve105.4–119.9scoalescenceicon·wireframe♪ bed_out
So wire one data source, ask one real question in words, save it to run weekly. Start with the one number you actually care about. The walkthrough, with the query we use, is at office dot temerarii dot xyz.
on-screen: Try it on your own numbers
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→Try it on your own numbers

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)

youtubeStop Staring at Dashboards: Let an Agent Read Your Numbers and Tell You What to Do You have data, and probably ten dashboards nobody reads. This walkthrough covers business intelligence the useful way: make an agent read your own numbers and tell you, in a sentence, what changed and what to do. The goal is an answer, not another chart you scroll past on a Monday. The method: - Point an agent at your real data. Give it the numbers, not a screenshot of the numbers. Wire an MCP server into your analytics, like Google Search Console or your database, and let the model pull the rows directly. Now it reads the live source, the same one your dashboard reads. - Ask the question in plain words. Instead of building a report, ask. Type: which pages lost the most search clicks this month, and why. The agent queries the data, does the math, and answers in plain language. You skipped the part where a human exports to a spreadsheet and color-codes it for an hour. - Let it write the next step, not just the finding. Ask the agent to turn its own finding into an action, like rewrite the three weakest page titles using the queries people actually searched. It reads your real query data and drafts the fix. The insight and the next move come out in the same run. - Save the question, run it weekly. Write that question into a workflow file so it runs every Monday on its own. Same query, fresh data, a short answer in your inbox before you open your laptop. A recurring answer is the only kind anyone reads. Proof without a figure: we point this at our own site's search data and let the model tell us which pages to fix next, in public. The pages you will see were edited based on what real searchers typed, not on a hunch in a meeting. Your next step: wire one data source, ask one real question in words, save it to run weekly. Start with the one number you actually care about. The walkthrough, with the query we use, is at office.temerarii.xyz. Keywords: business intelligence, MCP server, Google Search Console, plain language queries, weekly report, analytics automation.

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