Inputs Converge to One Result

Join separate inputs into one shared result with staged connectors. Support vertical elbows or horizontal curves without changing the convergence logic.

Prepare

Choose concise original content. Bundled generated images are reused automatically; supply local images only when replacing them. Presenter effects still require suitable footage. The sample values are illustrative. Read the included motion specification before replacing an essential relationship.

Generate

Run the installed Skill or use its bundled generator. Supply an output folder with --out, optional content JSON with --data, and a local video with --media for presenter effects. An optional --images JSON map replaces only its named slots; relative paths resolve beside that JSON file. Read references/image-policy.md before adapting visual material. The default style is white-yellow; alternatives are Charcoal & Lime (charcoal-lime), Paper & Orange (paper-orange), and Black & White (the compatible midnight-ice ID). Inter, Noto Sans SC and Smiley Sans are bundled locally. Chinese content uses Smiley Sans by default; pass --zh-font noto-sans-sc for the plainer Noto Sans SC. IBM Plex Mono is used for measurements and ordinal labels. Dark canvases have a fine grid.

Refine

Change content at its input, and motion in the effect source. Use semantic color tokens and keep type consistent. Rebuild the editable composition, run HyperFrames checks, and render the MP4. Review the beginning, key change and settled state before using it in a finished edit.

Inputs

• inputs: provide your own content or retain the illustrative default.
• result: provide your own content or retain the illustrative default.
• status: provide your own content or retain the illustrative default.
• orientation: provide your own content or retain the illustrative default.
• compactOnResolve: provide your own content or retain the illustrative default.
• resultInitiallyVisible: provide your own content or retain the illustrative default.

The output directory contains the complete local composition and build record. Generated image defaults and provenance are included in the installable Skill; preview videos and review reports are delivered separately.