/vox · Vox documentary collage
Best for: stories with numbers: how the business started, a case, myth vs fact, 'how X really works'.
Viral estimate: 7/10: strong curiosity and comments; needs a clear number or mystery to carry it. A judgement until their own posts in this style come back; /analytics updates it.
The style
- One archival tabletop: tan paper, ink #1A1A1A, ONE hot red #D62E1F, mustard #D9A441 for folders. The
"Fa"face at 900 on paper strips. Halftone only on real photos. - Face windows: their live face in black and white (
#face{filter:grayscale(1) contrast(1.22)}; a blend layer cannot reach the video), a light dot screen and a burned edge, a taped label. - B-roll: paper cut-outs stacking, rubber stamps slamming (scale 2 → 1 in 0.2 s), index cards pinned with brass pins and tied with red string, a camera pan across one wide board (right to left), manila folders, red stat cards counting up, newspaper headlines.
- Paper never fades in: it slides, springs (back.out), drops or is stamped. Hard cuts or a paper wipe between scenes.
- Captions: each word on its own paper strip, the spoken word turns red.
- Never generated faces or fake archive photos: only their real footage and real screenshots.
The style keeps its own palette on purpose (that is what makes it a style). The creator's brand enters through the "Fa"
font, assets/brand.js, the captions and the cover.
How to run it
Follow playbook/STYLE-REEL.md end to end with --style vox. The reference implementation of every scene type in this
style is tools/style-reel/styles/vox.js (scene ids B1–B6 are the same in every style; the template's demo is the worked
example). Rewrite the scenes for the new content in this style's grammar; never paste the demo text.
Before anything is shown: qa.py --plan with zero FAIL and the 3 fps sheet eye pass (playbook/MOTION-QA.md). Facts on
screen come only from the source (the owner's facts table, their own pages, a cited source).