One source in. Four kinds of artifact out.

Most teams do not have a content problem. They have a distribution problem: the thinking already exists as a webinar, a support doc, a customer call, a 2,000-word post — and it only ever ships once, on one network.

The engine runs in two stages, and the split is deliberate. An angles pass reads the whole source and names the separate arguments inside it. Then each angle gets its own model run. If one run fails you lose that angle, not the set — which is the difference between re-running one expansion and re-running the upload.

Each angle comes back as four things, and all four reach you: drafts written natively per network, slide-by-slide carousel copy, a scene-directed short-video script, and a Pin title and description. Slide copy opens in the carousel builder, a scene script opens in the video studio, the Pin opens in the composer. Nothing is generated and then quietly dropped.

Nothing is a copy-paste of the same caption either. Each output respects that network’s length, tone, hashtag convention and truncation point — read from the same table that drives the previews — so the first sentence still carries the post after the feed clips it.