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-W50-Wedkind longformweek W50date 2026-12-16campaign longform-youtubepillar it_devbeat asset videoduration 167.1sground 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-W50-Wed good|bad)
⚠ 1 flag(s) — not yet ship-ready: custom_element_dup · 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 · 167.1s · comp_id + rendered still + tier + the script

#Layer (comp_id · still · tier)BeatTimecodeMotionLogoAudioVO / on-screen / caption
1s1
matches intent
shared field
signature-3d
open0–23.8sspatial-parallaxicon·ember-fill♪ bed_in
Today is data and analytics for marketers, and the goal is to make your team stop waiting on reports. By the end, each person will pull a real answer from your own numbers, by asking in plain English. No spreadsheets-of-doom, no analyst bottleneck. Open the workspace. The first thing we fix is the question, because most data work fails before the data even shows up.
on-screen: Read your own data today
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→Read your own data today
2s2
matches intent
NumberedList
template
teach23.8–46.8skinetic-buildicon·color♪ node_lock
First move: write a real question, not a vague one. Not how are we doing. Instead: which channel brought the most new visitors last month. We have your team write three sharp questions about their own work. A sharp question has a time window, a thing to count, and a thing to group by. Get that right and the data part is easy.
on-screen: Ask a sharp question
expected on screen: white ground · a NumberedList panel over a dimmed Signal Field · Faber leads · node-graph · kinetic-build · icon·color logo · caption bottom-left
spec (the prompt): comp_id NumberedListvisual node-graphshape boxground whitetreatment colormotion kinetic-buildpower morphinstrument morph+laser→Ask a sharp questioncurate items, nodes
3s3
matches intent
ChecklistCard
template
teach46.8–70.6skinetic-buildicon·color♪ node_lock
Second move: connect the source. We wire an MCP server into Google Search Console and your analytics, so the agent can pull your real query data directly. Your team asks, what searches brought people to us, and the model pulls it and shows the table. They see their own words become a live answer. The data now sits next to the person making the call.
on-screen: Wire the model to Search Console
expected on screen: white ground · a ChecklistCard panel over a dimmed Signal Field · Faber leads · node-graph · kinetic-build · icon·color logo · caption bottom-left
spec (the prompt): comp_id ChecklistCardvisual node-graphshape boxground whitetreatment colormotion kinetic-buildpower morphinstrument morph+laser→Wire the model to Search Consolecurate items, nodes
4s4
first render · fix pending
BlueprintGrid
templatecustom_element_dup
teach70.6–95.8skinetic-buildicon·color♪ node_lock
Third move: trust but verify. The agent does not just hand you a number, it shows the query it wrote to get it. We teach your team to read that query and sanity-check it against one row they know by hand. This is the habit that keeps you safe. A confident wrong number is worse than no number, so you always check the math before you act on it.
on-screen: Make the model show its work
expected on screen: white ground · a BlueprintGrid panel over a dimmed Signal Field · Faber leads · node-graph · kinetic-build · icon·color logo · caption bottom-left
spec (the prompt): comp_id BlueprintGridvisual node-graphshape boxground whitetreatment colormotion kinetic-buildpower morphinstrument morph+laser→Make the model show its workcurate nodes, rows
5s5
matches intent
LegacyCard
template
teach95.8–118.8skinetic-buildicon·color♪ node_lock
Fourth move: act on it. Once the model can read your search data, we have it rewrite a weak page using the exact words people searched for. Not made-up keywords, your real ones. The team picks one underperforming page and lets the model redraft the heading and intro from the query data, then edits. Data goes straight into a change you can ship.
on-screen: Let the data rewrite the page
expected on screen: white ground · a LegacyCard panel over a dimmed Signal Field · Faber leads · node-graph · kinetic-build · icon·color logo · caption bottom-left
spec (the prompt): comp_id LegacyCardvisual node-graphshape boxground whitetreatment colormotion kinetic-buildpower morphinstrument morph+laser→Let the data rewrite the pagecurate nodes
6s6
matches intent
ComparisonTable
template
proof118.8–142.6sreceipts-counticon·color♪ node_lock
Honest proof: The Big T-M does this on itself. Our content engine has a scorer that reads what actually performed and learns from it, good and bad, so the next round is sharper. We do not guess what worked. We pull our own data and let it teach the machine. That loop is exactly what your team just built today, on your numbers, not ours.
on-screen: We learn from our own numbers
expected on screen: white ground · a ComparisonTable panel over a dimmed Signal Field · Faber leads · receipts · receipts-count · icon·color logo · caption bottom-left
spec (the prompt): comp_id ComparisonTablevisual receiptsshape boxground whitetreatment colormotion receipts-countpower receiptsinstrument spotlight→We learn from our own numberscurate colA, colB, rows, statsLabels
7s7
matches intent
shared field
signature-3d
resolve142.6–167.1scoalescenceicon·ember-fill♪ bed_out
The takeaway: analytics is not a monthly report, it is a question you can ask any time and a number you trust enough to act on. Your team now owns that loop. They can pull, check, and use their own data without waiting on anyone. See the live system at office dot temerarii dot xyz, and tomorrow we build the agent that runs these moves for you.
on-screen: Numbers where decisions live
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→Numbers where decisions live

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)

youtubeMarketing analytics: ask your own data a plain question, get an answer The goal is to make your team stop waiting on reports. By the end, each person pulls a real answer from your own numbers by asking in plain English. No spreadsheets of doom, no analyst bottleneck. The moves: - Ask a sharp question. Not how are we doing. Instead: which channel brought the most new visitors last month. A sharp question has a time window, a thing to count, and a thing to group by. Get that right and the data part is easy. - Wire the model to Search Console. Connect an MCP server into Google Search Console and your analytics so the agent pulls your real query data directly. Ask what searches brought people to us, and the model pulls it and shows the table. The data sits next to the person making the call. - Make the model show its work. The agent does not just hand you a number; it shows the query it wrote to get it. Read that query and sanity-check it against one row you know by hand. A confident wrong number is worse than no number, so check the math before you act on it. - Let the data rewrite the page. Once the model can read your search data, have it rewrite a weak page using the exact words people searched for, not made-up keywords, your real ones. Pick one underperforming page, let the model redraft the heading and intro from the query data, then edit. The honest proof: our content engine has a scorer that reads what actually performed and learns from it, good and bad, so the next round is sharper. We do not guess what worked; we pull our own data and let it teach the machine. Analytics is not a monthly report. It is a question you can ask any time and a number you trust enough to act on. See the live system at office.temerarii.xyz Keywords: marketing analytics, Google Search Console, MCP analytics, ask your data in plain English, verify the query, data-driven page edits

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