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-W35-Satkind longformweek W35date 2026-09-05campaign longform-youtubepillar staff_trainingbeat asset videoduration 161.3sground 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-W35-Sat 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 · 161.3s · comp_id + rendered still + tier + the script

#Layer (comp_id · still · tier)BeatTimecodeMotionLogoAudioVO / on-screen / caption
1s1
matches intent
shared field
signature-3d
open0–25.9sspatial-parallaxicon·3d-extrude♪ bed_in
We close the week the way we should: choosing an automation platform with no copy-paste, with accessibility in mind. A system that only works for people exactly like you is half built. Accessibility is not a favor you add at the end. It is the part that decides who gets to use the thing at all. We will teach plain moves that make your work usable by more people, on purpose.
on-screen: Build it so everyone can use it
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→Build it so everyone can use it
2s2
matches intent
NumberedList
template
teach25.9–52.2skinetic-buildicon·color♪ node_lock
First move. Every image needs alt text, a short written description a screen reader can speak aloud to someone who cannot see it. You can have a model draft the alt text from the image, then you fix the one detail it got wrong. The work is given away to the machine, the judgment stays with you. An image with no alt text is a door with no handle for some readers.
on-screen: Write alt text the machine drafts
expected on screen: white ground · a NumberedList panel over a dimmed Signal Field · Magister leads · node-graph · kinetic-build · icon·color logo · caption bottom-left
spec (the prompt): comp_id NumberedListvisual node-graphshape knotground whitetreatment colormotion kinetic-buildpower morphinstrument morph+laser→Write alt text the machine draftscurate items, nodes
3s3
matches intent
ChecklistCard
template
teach52.2–73.1skinetic-buildicon·color♪ node_lock
Second, make the text readable. There is a real measure for this, a contrast ratio between text and background, and tools that score it for you. Run the check, fix anything that fails, do not eyeball it. Light gray on white looks fine to you and disappears for someone else. A number settles it where taste cannot.
on-screen: Check contrast with a number
expected on screen: white ground · a ChecklistCard panel over a dimmed Signal Field · Magister leads · node-graph · kinetic-build · icon·color logo · caption bottom-left
spec (the prompt): comp_id ChecklistCardvisual node-graphshape knotground whitetreatment colormotion kinetic-buildpower morphinstrument morph+laser→Check contrast with a numbercurate items, nodes
4s4
matches intent
CodeWindow
template
teach73.1–95.8skinetic-buildicon·color♪ node_lock
Third, caption your videos. Run the audio through a speech-to-text model to draft the captions, then read them once for the names it mangled. Now your video works with the sound off, in a loud room, and for someone who is deaf. We caption ours for exactly this reason. The draft is free from the machine; the correction is the human part.
on-screen: Caption every video you ship
expected on screen: white ground · a CodeWindow panel over a dimmed Signal Field · Magister leads · node-graph · kinetic-build · icon·color logo · caption bottom-left
spec (the prompt): comp_id CodeWindowvisual node-graphshape knotground whitetreatment colormotion kinetic-buildpower morphinstrument morph+laser→Caption every video you shipcurate codeLines, nodes, windowTitle
5s5
matches intent
DiffCard
template
teach95.8–118.8skinetic-buildicon·color♪ node_lock
Our proof: the work at office dot temerarii dot xyz is built with these checks in the loop, not bolted on after. No compliance score to wave at you. The honest self-proof is that the captions, the alt text, and the contrast checks are part of how we ship, and the result is sitting in public for anyone, including assistive tools, to read.
on-screen: We ship it accessible, in public
expected on screen: white ground · a DiffCard panel over a dimmed Signal Field · Magister leads · node-graph · kinetic-build · icon·color logo · caption bottom-left
spec (the prompt): comp_id DiffCardvisual node-graphshape knotground whitetreatment colormotion kinetic-buildpower morphinstrument morph+laser→We ship it accessible, in publiccurate lines, nodes
6s6
matches intent
TerminalRun
template
proof118.8–138.3sreceipts-counticon·color♪ node_lock
The mindset: bake accessibility into the workflow so it is not a chore someone forgets. Let the model draft alt text and captions on every run, run the contrast check automatically, and make a human approve. When it is a step in the machine, it happens every time, instead of when someone remembers.
on-screen: Make it a step, not a chore
expected on screen: white ground · a TerminalRun panel over a dimmed Signal Field · Magister leads · receipts · receipts-count · icon·color logo · caption bottom-left
spec (the prompt): comp_id TerminalRunvisual receiptsshape knotground whitetreatment colormotion receipts-countpower receiptsinstrument spotlight→Make it a step, not a chorecurate statsLabels
7s7
matches intent
shared field
signature-3d
resolve138.3–161.3scoalescenceicon·3d-extrude♪ bed_out
So the takeaway for the week: a platform you choose well still has to serve everyone. Draft alt text and captions with a model, check contrast with a number, and put it all inside the workflow. Then go see the standard we hold ourselves to at office dot temerarii dot xyz. Your next step is one image, one honest line of alt text.
on-screen: See the standard at office.temerarii.xyz
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→See the standard at office.temerarii.xyz

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)

youtubeBuild it so everyone can use it: alt text, contrast scores, and captions baked into the workflow We close the week on choosing an automation platform with no copy-paste, with accessibility in mind. A system that only works for people exactly like you is half built. Accessibility is not a favor you add at the end. It decides who gets to use the thing at all. What you learn: - Every image needs alt text, a short written description a screen reader can speak aloud. Have a model draft it from the image, then fix the one detail it got wrong. An image with no alt text is a door with no handle for some readers. - Make the text readable. There is a real measure, a contrast ratio between text and background, and tools that score it. Run the check, fix anything that fails, do not eyeball it. A number settles it where taste cannot. - Caption your videos. Run the audio through a speech-to-text model to draft the captions, then read them once for the names it mangled. Now the video works with sound off, in a loud room, and for someone who is deaf. - Bake all of it into the workflow so it is not a chore someone forgets. Let the model draft on every run, run the contrast check automatically, and make a human approve. Our proof: the work at office.temerarii.xyz is built with these checks in the loop, not bolted on after. No compliance score to wave at you, just captions, alt text, and contrast checks that are part of how we ship, sitting in public for anyone, including assistive tools, to read. See the standard at office.temerarii.xyz. Your next step is one image, one honest line of alt text. Keywords: web accessibility, alt text, contrast ratio, video captions, screen readers, WCAG

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