Brand Systems
ShipTeaser
Early-stage companies ship constantly and announce badly, because a launch video costs an agency and two weeks they do not have. The design problem was never one video. It was the set of rules that lets one system produce a good one every time.
- Product design
- Template system
- Motion system
- Built with Claude Code
01 — The problem
Zoom out before zooming in
A founder cannot brief a video. They can paste a link. Everything downstream of that had to be a decision the system makes on their behalf, which means every decision had to be written down as a rule instead of made by taste on the day.
02 — The system
From one URL to a cut video
Four stages, each with a defined contract. The interesting constraint is stage three: the template has to be rigid enough to guarantee a result and loose enough that two customers never get the same film.
01 Input
A single product URL. The only thing a founder has time to give you.
02 Extraction
Name, value proposition, feature set, screenshots, brand colours and type read off the page.
03 Template
A fixed frame structure with typed slots. Content changes, timing and hierarchy do not.
04 Output
A cut video, sized for the feed it will be posted to.
03 — The template
Typed slots, fixed timing
The template is specified the way a grid is specified: named regions, fixed durations, a strict hierarchy. Content is variable, structure is not. That is what makes the output predictable without making it identical.
04 — The outputs
One system, many films
Because the structure is fixed, the system can render a different film per product without anyone art-directing it. The same six regions absorb a developer tool, a consumer app or a marketplace and still cut to length.
Dev tool
Consumer app
Marketplace
API product
Mobile app
Hardware
Structure shown as a spec, not as customer work.