Feature announcement · 20s · 16:9
One feature, one spotlight
This finished sample was made from captured screens and page copy of imagerework.com, one of our own products. It is not a live generator result. Silent cut — captions carry it.
Current generator access is not open yet. The workbench reads public page text, turns reviewable facts into an editable script and timed scenes using animated typography, cards and simple shapes. It does not capture or recreate your live UI.
This prompt records how the existing video above was made. It is not a capability list or a prompt that the current workbench runs.
Look: developer-dark launch film — near-black background (#0b0b0d) with a faint grid, a sans display face plus monospace accents, the product's own brand color as the single accent. Show the real screenshots inside a clean browser frame whose address bar shows the real domain; move a 'camera' across them with smooth pans and zooms, and point at real elements with a cursor or highlight box. This is a single-feature spotlight, not a full tour. Pick the 'mark only the areas that still need attention / refine up to five marked areas' capability that the page describes. 0–3s a bold kicker line with the feature name in the page's own words; 3–12s zoom into the screenshot region that shows this step, draw numbered highlight rings that build up one at a time, with a short caption per ring taken from the page; 12–16s one line explaining why it matters, worded from the page; 16–20s CTA + domain. Do NOT use the word 'new', version numbers or dates — the page does not state them.
// planned text workflow
- 01Enter a public URL and name one feature to highlight; the planned reader uses page text only.
- 02Review the source facts and confirm the feature wording.
- 03Shape the message into timed captions, cards and graphic callouts; no screenshot region is generated.
- 04Rendering is closed. The prompt above documents the finished sample, not current generator behavior.
// inspiration
Technique inspired by @leodev on X. Their video and prompt stay theirs — we don't host or reuse them. This sample, prompt and steps are ours.
Try the sample walkthrough with mock facts and a pre-rendered clip, or join the waitlist for text-based rendering when it opens.