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 topic-et-aikind topicweek —date —campaign emerging-tech-pillarpillar emerging_techbeat —asset videoduration 86.7sground blackscenes 14

Checklist the per-video bar — engine/sim

97.7/100
plain languagevo coverageno dead airuniquenesscaption fitcompletenesscleanliness
quantitative quality · weights learn from your reviews (engine.sim.memory review topic-et-ai good|bad)
⚠ 2 flag(s) — not yet ship-ready: copy_generictoo_complex · see docs/strategy/VIDEO-CHECKLIST.md

Composition comp · template family · expected output

composition SceneReelfamily / template SceneReel
9:16 Reelrendered1:1 Squarepending16:9 Widepending9:16 4Kpending1:1 4Kpending16:9 4KpendingGIF (SMS)pending
▶ open rendered mp4
expected output: 1/7 rendered — same matrix the /media preview surfaces for this asset.

Composition layer × scene 14 scenes · 86.7s · comp_id + rendered still + tier + the script

#Layer (comp_id · still · tier)BeatTimecodeMotionLogoAudioVO / on-screen / caption
1s1
matches intent
shared field
signature-3d		open0–6.6sspatial-parallaxicon·ember-fill♪ —
You have used the chatbot. Now let us show you what artificial intelligence does when you wire it into real work.
on-screen: Past the chatbot
expected on screen: black ground · cone hero in the shared Signal Field · Augur leads · node-graph · spatial-parallax · icon·ember-fill logo · caption bottom-left
spec (the prompt): comp_id shared Signal Fieldvisual node-graphshape coneground blacktreatment ember-fillmotion spatial-parallaxpower summoninstrument summon→three-mark
2s2
matches intent
FlowSchematic
template		legacy6.6–12.8scrossfade-8ficon·wireframe♪ —
The old way was pasting into a chat window and pasting the answer back, a smart intern with no hands.
on-screen: The copy-paste assistant
expected on screen: black ground · a FlowSchematic panel over a dimmed Signal Field · Augur leads · crossfade-8f · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id FlowSchematicshape coneground blacktreatment wireframemotion crossfade-8fpower morphinstrument decay→three-symboliccurate stages
3s3
matches intent
NumberedList
template		teach12.8–19.9skinetic-buildicon·wireframe♪ —
We give the model tools, real functions it can call, so it can read a file or send a request, not just
on-screen: Step 1: give it tools
expected on screen: black ground · a NumberedList panel over a dimmed Signal Field · Augur leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id NumberedListvisual node-graphshape coneground blacktreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→html-in-canvas-panelcurate items, nodes
4s4
matches intent
ChecklistCard
template		teach19.9–26.7skinetic-buildicon·wireframe♪ —
Each agent gets one well-bounded job and the tools for exactly that, because a vague agent fails vaguely.
on-screen: Step 2: a clear job
expected on screen: black ground · a ChecklistCard panel over a dimmed Signal Field · Augur leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id ChecklistCardvisual node-graphshape coneground blacktreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→three-diagramcurate items, nodes
5s5
matches intent
LogStream
template		teach26.7–32.4skinetic-buildicon·wireframe♪ —
We make it answer in a strict shape, so the next step gets clean data instead of parsing a paragraph.
on-screen: Step 3: structured output
expected on screen: black ground · a LogStream panel over a dimmed Signal Field · Augur leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id LogStreamvisual node-graphshape coneground blacktreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→html-in-canvas-codecurate nodes, rows
6s6
matches intent
JsonDiff
template		teach32.4–38.699999999999996skinetic-buildicon·wireframe♪ —
We feed it your real documents at answer time, so it cites your truth instead of inventing a plausible one.
on-screen: Step 4: ground it in your docs
expected on screen: black ground · a JsonDiff panel over a dimmed Signal Field · Augur leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id JsonDiffvisual node-graphshape coneground blacktreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→three-flowcurate fileName, lines, nodes
7s7
matches intent
SchematicCard
template		teach38.7–44.300000000000004skinetic-buildicon·wireframe♪ —
We cache the parts of the prompt that never change, so it runs faster and costs a fraction per
on-screen: Step 5: cache the context
expected on screen: black ground · a SchematicCard panel over a dimmed Signal Field · Augur leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id SchematicCardvisual node-graphshape coneground blacktreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→html-in-canvas-panelcurate nodes
8s8
matches intent
CodeWindow
template		teach44.3–50.4skinetic-buildicon·wireframe♪ —
A second pass reviews the first, because the cheapest way to catch a model's mistake is another model.
on-screen: Step 6: let it check itself
expected on screen: black ground · a CodeWindow panel over a dimmed Signal Field · Augur leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id CodeWindowvisual node-graphshape coneground blacktreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→three-diagramcurate codeLines, nodes, windowTitle
9s9
matches intent
StackTrace
template		teach50.4–56.5skinetic-buildicon·wireframe♪ —
Anything irreversible waits for a person to approve, so the AI proposes and a human disposes.
on-screen: Step 7: keep a human gate
expected on screen: black ground · a StackTrace panel over a dimmed Signal Field · Augur leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id StackTracevisual node-graphshape coneground blacktreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→html-in-canvas-codecurate errMsg, errType, fix, frames, nodes
10s10
matches intent
ProcessFlow
template		teach56.5–62.3skinetic-buildicon·wireframe♪ —
Every decision logs what it saw and why, so when it is wrong you can read the trail, not
on-screen: Step 8: log the reasoning
expected on screen: black ground · a ProcessFlow panel over a dimmed Signal Field · Augur leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id ProcessFlowvisual node-graphshape coneground blacktreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→three-flowcurate nodes, steps
11s11
matches intent
AnnotatedDiagram
template		teach62.3–67.7skinetic-buildicon·wireframe♪ —
We grade the agent on the real result it produced, not on how confident it sounded doing it.
on-screen: Step 9: measure the outcome
expected on screen: black ground · a AnnotatedDiagram panel over a dimmed Signal Field · Augur leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id AnnotatedDiagramvisual node-graphshape coneground blacktreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→html-in-canvas-panelcurate callouts, nodes
12s12
matches intent
RankList
template		proof67.7–75.10000000000001sreceipts-counticon·wireframe♪ —
That is AI with tools, a job, guardrails, and a log, an agent that finishes tasks instead of describing
on-screen: It actually did the work
expected on screen: black ground · a RankList panel over a dimmed Signal Field · Augur leads · receipts · receipts-count · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id RankListvisual receiptsshape coneground blacktreatment wireframemotion receipts-countpower receiptsinstrument spotlight→html-in-canvas-receiptscurate rows, statsLabels
13s13
matches intent
shared field
signature-3d		futurist75.1–80.6sspatial-parallaxicon·wireframe♪ —
When agents can do bounded work safely, your people move up to the work that needs a soul.
on-screen: What this unlocks
expected on screen: black ground · cone hero in the shared Signal Field · Augur leads · spatial-parallax · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id shared Signal Fieldshape coneground blacktreatment wireframemotion spatial-parallaxpower spatial-parallaxinstrument laser-fire→three-forward
14s14
matches intent
shared field
signature-3d		resolve80.6–86.69999999999999scoalescenceicon·ember-fill♪ swell
Pick one task you wish ran itself. We will build the agent that does it. temerarii.xyz.
on-screen: Pick one task
expected on screen: black ground · cone hero in the shared Signal Field · Augur leads · coalescence · coalescence · icon·ember-fill logo · caption bottom-left
spec (the prompt): comp_id shared Signal Fieldvisual coalescenceshape coneground blacktreatment ember-fillmotion coalescencepower coalescenceinstrument coalescence→three-coalescence

Format stack 3 aspects · same scenes[], re-cropped

9:16
1080×1920
Stories · TikTok · YouTube Shorts · Reels
1:1
1080×1080
LinkedIn · Facebook · Instagram
16:9
1920×1080
X/Twitter · YouTube · LinkedIn video

Channels 12 destinations

LinkedInX/TwitterYouTubeInstagramFacebookThreadsTikTokPinterestBlueskyEmailSMSBlog

Social captions supplemental published copy · per channel (comp_id level)

tiktokYou've used the chatbot. Here's what AI does when you wire it into real work. (AI-assisted) Give the model real tools it can call, one clear job, and make it answer in a strict shape. Ground it in your own docs, let a second pass check the first, and keep a human gate on anything you can't undo. That's an agent that finishes tasks. Pick one task that should run itself. temerarii.xyz
instagramYou've used the chatbot. Here's AI wired into real work. Give the model real tools to call. One clear job per agent. Make it answer in a strict shape. Ground it in your own docs. Keep a human gate on the irreversible. Pick one task that should run itself. #aiagents #automation #appliedai #mcp
linkedinA chatbot is a smart intern with no hands. An agent has hands and a job. You have used the chat window. Here is what AI does when you wire it into real work: 1. Give the model real tools, functions it can call to read a file or send a request. 2. Give each agent one well-bounded job and the tools for exactly that. 3. Make it answer in a strict shape, so the next step gets clean data, not a paragraph. 4. Feed it your real documents at answer time, so it cites your truth instead of inventing one. 5. Cache the parts of the prompt that never change, so it runs faster and cheaper. 6. Let a second pass review the first. 7. Keep a human gate on anything irreversible. 8. Log what it saw and why, so when it is wrong you can read the trail. The takeaway: pick one task you wish ran itself, and build the agent that does it. temerarii.xyz
xA chatbot is a smart intern with no hands. An agent has hands and a job. Give the model real tools, one bounded task, strict output, your docs to cite, and a human gate on anything irreversible. Pick one task that should run itself.
facebookYou've used the chatbot. Here's what AI does when you wire it into real work. Give the model real tools it can call, one well-bounded job, and make it answer in a strict shape. Ground it in your own documents so it cites your truth, let a second pass check the first, and keep a human gate on anything you can't undo. That's an agent that finishes tasks. Pick one task you wish ran itself, and we'll build it. temerarii.xyz
threadsYou've used the chatbot. Here's AI wired into real work. Give the model real tools it can call. One bounded job per agent. A strict output shape. Ground it in your own docs. Let a second pass check the first. Keep a human gate on anything irreversible. Pick one task that should run itself.
pinterestApplied AI agent method: how to wire a language model into real work. Give the model real tools it can call, one well-bounded job, and strict structured output. Ground it in your own documents so it cites your truth, cache the static prompt, add a self-check pass, and keep a human gate on irreversible actions. Build AI agents that finish tasks.
blueskyA chatbot is a smart intern with no hands. An agent has hands and a job. Give the model real tools, one bounded task, strict output, your docs to cite, and a human gate on the irreversible. Pick one task that should run itself. temerarii.xyz
youtubeBeyond the Chatbot: How to Build an AI Agent That Finishes Real Work You have used the chatbot. Now here is what artificial intelligence does when you wire it into real work. The old way was pasting into a chat window and pasting the answer back, a smart intern with no hands. Here is the method: give the model real tools, functions it can call so it can read a file or send a request; give each agent one well-bounded job and the tools for exactly that; make it answer in a strict shape so the next step gets clean data; feed it your real documents at answer time so it cites your truth instead of inventing one; cache the parts of the prompt that never change so it runs faster and cheaper; let a second pass review the first; keep a human gate on anything irreversible; log what it saw and why; and grade it on the real result, not on how confident it sounded. When agents can do bounded work safely, your people move up to the work that needs a soul. Pick one task you wish ran itself, and we will build the agent that does it. temerarii.xyz

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

→ Wireframe · scenes→ Calendar week→ Media (audience preview)→ Storyboard hub→ Day view→ Channel view