Checklist the per-video bar — engine/sim
100.0/100
quantitative quality · weights learn from your reviews (
engine.sim.memory review longform-W42-Mon good|bad)✓ all static checks pass — one-focal/scene · tier-by-beat · one-track caption · colorway · cast+shape correct · no banned/fabricated. (audio + visual tiers verify on the rendered finals — Phase 2)
Composition comp · template family · expected output
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 6 scenes · 149.3s · 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 | 0–24.8s | spatial-parallax | icon·white-knockout | ♪ bed_in | Today is about photography for brands that cannot stop and run a studio shoot every week. The promise of this week is proof you are real, and photos are a big part of that. You will learn how to build a matching set of brand images using an AI model and an API, keep them consistent, and still leave room for the real camera when you have it. on-screen: Photos when you cannot shoot |
expected on screen: red ground · icosa hero in the shared Signal Field · Signum leads · node-graph · spatial-parallax · icon·white-knockout logo · caption bottom-left spec (the prompt): comp_id shared Signal Fieldvisual node-graphshape icosaground redtreatment white-knockoutmotion spatial-parallaxpower summoninstrument summon→Photos when you cannot shoot | |||||||
| 2 |  matches intent NumberedList template | teach | 24.8–50.7s | kinetic-build | icon·white-knockout | ♪ node_lock | First step: write a style block once. Pick your model, we use Flux through the Replicate API, and write a short fixed description, the lighting, the color, the mood, in plain words. Save it. Every image you generate starts with that same block, so shot one and shot fifty look like they came from the same camera on the same day. The style lives in a file, not in your head. on-screen: Lock a style, not a vibe |
expected on screen: red ground · a NumberedList panel over a dimmed Signal Field · Signum leads · node-graph · kinetic-build · icon·white-knockout logo · caption bottom-left spec (the prompt): comp_id NumberedListvisual node-graphshape icosaground redtreatment white-knockoutmotion kinetic-buildpower morphinstrument morph+laser→Lock a style, not a vibecurate items, nodes | |||||||
| 3 |  matches intent ChecklistCard template | teach | 50.7–77.30000000000001s | kinetic-build | icon·liquid-chrome | ♪ node_lock | Second step: use a seed number. A seed is just a starting point the model uses to draw. Same seed plus same prompt gives you nearly the same image again, so you can tweak one small thing instead of rolling the dice every time. We log the seed for every keeper. When a client says make it warmer, we change one word and keep the seed, and the rest of the shot holds. on-screen: Seed it so it repeats |
expected on screen: red ground · a ChecklistCard panel over a dimmed Signal Field · Signum leads · node-graph · kinetic-build · icon·liquid-chrome logo · caption bottom-left spec (the prompt): comp_id ChecklistCardvisual node-graphshape icosaground redtreatment liquid-chromemotion kinetic-buildpower morphinstrument morph+laser→Seed it so it repeatscurate items, nodes | |||||||
| 4 |  matches intent JsonDiff template | teach | 77.3–101.8s | kinetic-build | icon·white-knockout | ♪ node_lock | Third step: do not throw away the camera. When you do shoot, drop those files in the same folder as the generated ones, named the same way. The pipeline treats them as equals. Generated images fill the gaps so the grid is never empty, real photos carry the moments that have to be true. Nobody scrolling can tell where the seam is, and that is the point. on-screen: Real photos go in the same folder |
expected on screen: red ground · a JsonDiff panel over a dimmed Signal Field · Signum leads · node-graph · kinetic-build · icon·white-knockout logo · caption bottom-left spec (the prompt): comp_id JsonDiffvisual node-graphshape icosaground redtreatment white-knockoutmotion kinetic-buildpower morphinstrument morph+laser→Real photos go in the same foldercurate fileName, lines, nodes | |||||||
| 5 |  matches intent RankList template | proof | 101.8–126.3s | receipts-count | icon·liquid-chrome | ♪ node_lock | Our own image library was built exactly like this, a locked style block and logged seeds, no weekly studio booking. We are not claiming it beats a top photographer on their best day. We are saying it gives a small brand a full, matching set instead of three stock photos and an excuse. That is an honest trade, and it is the one most businesses actually need. on-screen: We built our own grid this way |
expected on screen: red ground · a RankList panel over a dimmed Signal Field · Signum leads · receipts · receipts-count · icon·liquid-chrome logo · caption bottom-left spec (the prompt): comp_id RankListvisual receiptsshape icosaground redtreatment liquid-chromemotion receipts-countpower receiptsinstrument spotlight→We built our own grid this waycurate rows, statsLabels | |||||||
| 6 |  matches intent shared field signature-3d | resolve | 126.3–149.3s | coalescence | icon·white-knockout | ♪ bed_out | So, photography without the photoshoot bottleneck: one style block, a logged seed, and a folder where real and generated photos sit side by side. It is calm, it repeats, and it never leaves you with a blank page. You can see the whole image system running at office.temerarii.xyz. The next step is easy, write your style block today and generate three test shots. on-screen: Try it at office.temerarii.xyz |
expected on screen: red ground · icosa hero in the shared Signal Field · Signum leads · coalescence · coalescence · icon·white-knockout logo · caption bottom-left spec (the prompt): comp_id shared Signal Fieldvisual coalescenceshape icosaground redtreatment white-knockoutmotion coalescencepower coalescenceinstrument coalescence→Try it 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 build a matching brand photo set with AI when you cannot shoot
Not every brand can stop and run a studio shoot every week. This video shows how to build a matching set of brand images using an AI model and an API, keep them consistent, and still leave room for the real camera when you have it.
The method:
Lock a style, not a vibe. Write a style block once, naming the model, the lighting, the color, and the mood in plain words. We use Flux through the Replicate API. Save it, and every image you generate starts with that same block, so shot one and shot fifty look like they came from the same camera on the same day.
Seed it so it repeats. A seed is the starting point the model uses to draw. Same seed plus same prompt gives you nearly the same image again, so you can tweak one small thing instead of rolling the dice. Log the seed for every keeper. When a client says make it warmer, change one word and keep the seed, and the rest of the shot holds.
Put real photos in the same folder. When you do shoot, drop those files in the same folder as the generated ones, named the same way. The pipeline treats them as equals. Generated images fill the gaps so the grid is never empty, real photos carry the moments that have to be true.
This is an honest trade. It will not beat a top photographer on their best day, but it gives a small brand a full, matching set instead of three stock photos and an excuse.
See the whole image system at office.temerarii.xyz, then write your style block today and generate three test shots.
Keywords: AI photography, Flux, Replicate, brand photos, seed, image generation, content pipeline.