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-W42-Wedkind longformweek W42date 2026-10-21campaign longform-youtubepillar socialbeat asset videoduration 183.8sground blackscenes 8

Checklist the per-video bar — engine/sim

98.4/100
plain languagevo coverageno dead airuniquenesscaption fitcompletenesscleanliness
quantitative quality · weights learn from your reviews (engine.sim.memory review longform-W42-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 8 scenes · 183.8s · comp_id + rendered still + tier + the script

#Layer (comp_id · still · tier)BeatTimecodeMotionLogoAudioVO / on-screen / caption
1s1
matches intent
shared field
signature-3d
open0–22.3sspatial-parallaxicon·ember-fill♪ bed_in
Today is live streaming, the scariest format because you cannot edit it after. The week is about proving you are real, and nothing proves it like going live without a net. You will learn how to set up a stream where an AI assistant handles the parts that usually make a solo host crash, so you can just talk to people.
on-screen: Live streaming, run by a co-pilot
expected on screen: black ground · torus hero in the shared Signal Field · Nuntius leads · node-graph · spatial-parallax · icon·ember-fill logo · caption bottom-left
spec (the prompt): comp_id shared Signal Fieldvisual node-graphshape torusground blacktreatment ember-fillmotion spatial-parallaxpower summoninstrument summon→Live streaming, run by a co-pilot
2s2
matches intent
NumberedList
template
teach22.3–45.3skinetic-buildicon·wireframe♪ node_lock
First step: use OBS, the free open source streaming software. It captures your camera, your screen, and your overlays, and pushes one feed out to any platform. There is no subscription holding your stream hostage. We set up scenes in OBS once, a talking head scene, a screen share scene, a be right back scene, and switch between them with a key press.
on-screen: Open source software, no monthly fee
expected on screen: black ground · a NumberedList panel over a dimmed Signal Field · Nuntius leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id NumberedListvisual node-graphshape torusground blacktreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→Open source software, no monthly feecurate items, nodes
3s3
matches intent
ChecklistCard
template
teach45.3–68.3skinetic-buildicon·wireframe♪ node_lock
Second step: pipe your live chat into a model. The chat API for the platform sends every message to a small script, and an AI reads along, flagging the real questions and hiding the noise. Instead of squinting at a wall of comments while talking, you get a short list of what actually matters. The model is your producer whispering in your ear.
on-screen: Read the chat with a model
expected on screen: black ground · a ChecklistCard panel over a dimmed Signal Field · Nuntius leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id ChecklistCardvisual node-graphshape torusground blacktreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→Read the chat with a modelcurate items, nodes
4s4
matches intent
AnnotatedDiagram
template
teach68.3–91.3skinetic-buildicon·wireframe♪ node_lock
Third step: dynamic on screen text. We drive the lower third name bars and topic banners from a small data file, so when the topic changes the banner changes without you touching anything. The model can even write the next banner from what you just said. Your stream looks staffed by a graphics team when it is really one file and a model.
on-screen: Lower thirds that update themselves
expected on screen: black ground · a AnnotatedDiagram panel over a dimmed Signal Field · Nuntius leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id AnnotatedDiagramvisual node-graphshape torusground blacktreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→Lower thirds that update themselvescurate callouts, nodes
5s5
matches intent
ProcessFlow
template
teach91.3–112.19999999999999skinetic-buildicon·wireframe♪ node_lock
Fourth step: always record the local copy. OBS saves a clean high quality file while you stream. After you go offline, that file goes straight into the same pipeline as everything else, captions from Whisper, cuts in Remotion. One live hour becomes a dozen short clips with no extra filming. The live show pays for itself twice.
on-screen: Record once, cut it later for free
expected on screen: black ground · a ProcessFlow panel over a dimmed Signal Field · Nuntius leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id ProcessFlowvisual node-graphshape torusground blacktreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→Record once, cut it later for freecurate nodes, steps
6s6
matches intent
LogStream
template
teach112.2–134.5skinetic-buildicon·wireframe♪ node_lock
Fifth step: protect the people on the call. If you bring guests or read names, we run the recording through a step that can blur faces and mute private details before any clip goes public. The model marks the spots, you confirm. Going live in public should not mean leaking someone by accident. Honest proof never costs other people their privacy.
on-screen: Faces blurred, names kept private
expected on screen: black ground · a LogStream panel over a dimmed Signal Field · Nuntius leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id LogStreamvisual node-graphshape torusground blacktreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→Faces blurred, names kept privatecurate nodes, rows
7s7
first render · fix pending
ComparisonTable
templatecustom_element_dup
proof134.5–157.9sreceipts-counticon·wireframe♪ node_lock
We run our own studio out loud, the machine working where anyone can watch, no hidden cut. We are not claiming every stream is smooth, live never is. We are saying the setup, free software, a model reading chat, self updating graphics, and an auto saved recording, lets one person run a show that used to need a crew. That is the real win.
on-screen: We stream the studio in public
expected on screen: black ground · a ComparisonTable panel over a dimmed Signal Field · Nuntius leads · receipts · receipts-count · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id ComparisonTablevisual receiptsshape torusground blacktreatment wireframemotion receipts-countpower receiptsinstrument spotlight→We stream the studio in publiccurate colA, colB, rows, statsLabels
8s8
matches intent
shared field
signature-3d
resolve157.9–183.8scoalescenceicon·ember-fill♪ bed_out
So that is live streaming for a team of one: OBS for the feed, a model reading the chat, graphics that update from a file, a recording you reuse, and privacy handled on the way out. It turns the scariest format into a calm one. You can watch the studio run live at office.temerarii.xyz. The next step is tiny, install OBS and do a five minute test stream to no one.
on-screen: See it live at office.temerarii.xyz
expected on screen: black ground · torus hero in the shared Signal Field · Nuntius leads · coalescence · coalescence · icon·ember-fill logo · caption bottom-left
spec (the prompt): comp_id shared Signal Fieldvisual coalescenceshape torusground blacktreatment ember-fillmotion coalescencepower coalescenceinstrument coalescence→See it live 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)

youtubeHow to run a live stream solo with an AI co-pilot handling the hard parts Live streaming is the scariest format because you cannot edit it after. This video shows how to set up a stream where an AI assistant handles the parts that usually make a solo host crash, so you can just talk to people. The method: Use OBS, the free open source streaming software. It captures your camera, screen, and overlays and pushes one feed to any platform, with no subscription holding your stream hostage. Set up scenes once, a talking head scene, a screen share scene, a be right back scene, and switch with a key press. Read the chat with a model. The platform chat API sends every message to a small script, and a model reads along, flagging the real questions and hiding the noise. Instead of squinting at a wall of comments while talking, you get a short list of what matters. The model is your producer whispering in your ear. Drive the lower thirds from a data file, so when the topic changes the banner changes without you touching anything. The model can even write the next banner from what you just said. Your stream looks staffed by a graphics team when it is one file and a model. Always record the local copy. OBS saves a clean high quality file while you stream, and after you go offline it goes into the same pipeline as everything else, captions from Whisper, cuts in Remotion. One live hour becomes a dozen short clips. Protect the people on the call. Run the recording through a step that can blur faces and mute private details before any clip goes public. The model marks the spots, you confirm. Watch the studio run live at office.temerarii.xyz, then install OBS and do a five minute test stream to no one. Keywords: live streaming, OBS, solo streamer, AI assistant, Whisper, Remotion, chat moderation.

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