Checklist the per-video bar — engine/sim
78.3/100
quantitative quality · weights learn from your reviews (
engine.sim.memory review thread-build-log-W31-Wed good|bad)⚠ 7 flag(s) — not yet ship-ready: copy_genericlow_vo_coveragedead_airgeneric_scene · see docs/strategy/VIDEO-CHECKLIST.md
Composition comp · template family · expected output
9:16 Reelrendered1:1 Squarepending16:9 Widepending9:16 4Kpending1:1 4Kpending16:9 4KpendingGIF (SMS)pending
expected output: 1/7 rendered — same matrix the /media preview surfaces for this asset.
Composition layer × scene 6 scenes · 32.8s · comp_id + rendered still + tier + the script
| # | Layer (comp_id · still · tier) | Beat | Timecode | Motion | Logo | Audio | VO / on-screen / caption |
|---|---|---|---|---|---|---|---|
| 1 |  matches intent shared field signature-3d | open | 0–6.6s | spatial-parallax | icon·white-knockout | ♪ bed_in | Week nine. We shipped on-brand AI image generation this week, every asset in our exact palette. on-screen: Shipped: AI images on brand |
expected on screen: red ground · tetra hero in the shared Signal Field · Nexus leads · node-graph · spatial-parallax · icon·white-knockout logo · caption bottom-left spec (the prompt): comp_id shared Signal Fieldvisual node-graphshape tetraground redtreatment white-knockoutmotion spatial-parallaxpower summoninstrument summon→Shipped: AI images on brand | |||||||
| 2 |  matches intent BuildLog template | build | 6.6–11.6s | type-on | icon·white-knockout | ♪ node_lock | The old way was hunting stock photos that never matched the brand palette or the type. on-screen: Old way: stock photos, off-brand |
expected on screen: red ground · a BuildLog panel over a dimmed Signal Field · Nexus leads · code · type-on · icon·white-knockout logo · caption bottom-left spec (the prompt): comp_id BuildLogvisual codeshape tetraground redtreatment white-knockoutmotion type-onpower summoninstrument draw-on→Old way: stock photos, off-brandcurate codeLines, lines | |||||||
| 3 |  first render · fix pending DiffCard templatedead_air | diff | 11.6–16.6s | crossfade-8f | icon·liquid-chrome | ♪ node_lock | It broke a rule. It kept typing the brand name into the image, so we banned that in the prompt. on-screen: What broke: it typed the name |
expected on screen: red ground · a DiffCard panel over a dimmed Signal Field · Nexus leads · code · crossfade-8f · icon·liquid-chrome logo · caption bottom-left spec (the prompt): comp_id DiffCardvisual codeshape tetraground redtreatment liquid-chromemotion crossfade-8fpower morphinstrument morph→What broke: it typed the namecurate codeLines, lines | |||||||
| 4 |  matches intent PipelineMap template | step | 16.6–22.8s | kinetic-build | icon·white-knockout | ♪ node_lock | Now every image lands on-brand the first time. Next, we lock a recurring cast of characters. on-screen: Next: a reusable cast |
expected on screen: red ground · a PipelineMap panel over a dimmed Signal Field · Nexus leads · pipeline · kinetic-build · icon·white-knockout logo · caption bottom-left spec (the prompt): comp_id PipelineMapvisual pipelineshape tetraground redtreatment white-knockoutmotion kinetic-buildpower throwinstrument laser-trace→Next: a reusable castcurate stages, steps | |||||||
| 5 |  first render · fix pending StatScoreboard templatedead_airgeneric_scene | proof | 22.8–27.8s | receipts-count | icon·liquid-chrome | ♪ node_lock | Build Log |
expected on screen: red ground · a StatScoreboard panel over a dimmed Signal Field · Nexus leads · receipts · receipts-count · icon·liquid-chrome logo · caption bottom-left spec (the prompt): comp_id StatScoreboardvisual receiptsshape tetraground redtreatment liquid-chromemotion receipts-countpower receiptsinstrument spotlight→Build Logcurate pillar, stats, statsLabels | |||||||
| 6 |  first render · fix pending shared field signature-3ddead_airgeneric_scene | resolve | 27.8–32.8s | coalescence | icon·white-knockout | ♪ bed_out | Build Log |
expected on screen: red ground · tetra hero in the shared Signal Field · Nexus leads · coalescence · coalescence · icon·white-knockout logo · caption bottom-left spec (the prompt): comp_id shared Signal Fieldvisual coalescenceshape tetraground redtreatment white-knockoutmotion coalescencepower coalescenceinstrument coalescence→Build Log | |||||||
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 9 destinations
LinkedInX/TwitterYouTubeInstagramFacebookThreadsTikTokPinterestBluesky
Social captions supplemental published copy · per channel (comp_id level)
tiktokBuild log, week nine: we shipped on-brand AI image generation. Every asset in our exact palette. The old way was hunting stock that never matched. It broke once by typing the brand name into the image, so we banned that in the prompt. Next, a reusable cast. (AI-assisted)
instagramBuild Log · Week 9
Shipped: on-brand AI images.
Every asset, our exact palette.
What broke: it kept typing the brand name in.
Fix: banned that in the prompt.
Next: a reusable cast of characters.
#buildinpublic #aiimages #brand #designops #midjourney
linkedinBuild log, week nine. This week we shipped on-brand AI image generation: every asset now lands in our exact palette the first time. The old way was hunting stock photos that never matched the brand. What broke along the way: the model kept typing the brand name into the image, so we added that to the banned list in the prompt. Next up, we lock a recurring cast of characters so the visuals stay consistent across campaigns.
xBuild log wk9: shipped on-brand AI images, every asset in our exact palette. It kept typing the brand name into the image, so we banned that in the prompt. Next: a reusable cast. temerarii.xyz
facebookBuild log, week nine. We shipped on-brand AI image generation this week, so every asset lands in our exact palette. The old way was hunting stock photos that never matched. One thing broke: the model kept typing the brand name into the image, so we banned that in the prompt. Next, we lock a reusable cast of characters.
threadsBuild log, week nine. We shipped on-brand AI image generation, every asset in our exact palette. The old way was hunting stock that never matched. What broke: it kept typing the brand name into the image, so we banned that in the prompt. Next, a reusable cast of characters.
pinterestBuilding on-brand AI image generation, step by step. Week nine of a build log: every generated asset now lands in the exact brand palette, replacing off-brand stock photos. A common bug is the model typing the brand name into the image, fixed by adding it to a banned-words prompt. Next milestone: a reusable cast of characters.
blueskyBuild log wk9: shipped on-brand AI images, every asset in our exact palette. It kept typing the brand name into the image, so we banned that in the prompt. Next: a reusable cast.
youtubeTitle: Build Log Week 9: Shipping On-Brand AI Images (and the Bug We Hit)
This week we shipped on-brand AI image generation, so every asset lands in our exact palette instead of off-brand stock. We walk through the build, including the bug where the model kept typing the brand name into the image and how we banned it in the prompt. Next, we lock a reusable cast of characters.