Hyperframes

HyperFrames entry point

HyperFrames renders video from HTML — a composition is an HTML file whose DOM declares timing with data-* attributes, whose animation runtime is seekable, and whose media playback is owned by the framework. The full authoring contract lives in /hyperframes-core; read it before writing composition HTML.

1. Start from project state

Apply the first matching row; do not evaluate lower state rows:

If a fresh request does not identify the subject or input, ask what the video is about before routing. Check preferences and recipes before asking anything (references/intent-interview.md, step 1). A figma.com input or a named recipe changes intake, not routing — the interview's "Adapt orthogonal inputs" section handles both.

Keep the project's CLI current

A scaffolded project pins hyperframes@<version> in its package.json scripts so renders stay reproducible; the pin never advances on its own, and a pinned run of an older CLI prints no warning about it. When resuming a project whose scripts carry a pin, probe once before the first render-affecting command:

npx hyperframes@latest upgrade --project . --check

The probe is read-only and reports the pin against the latest release; keep the explicit . — on older CLI releases a bare --project followed by another flag consumes that flag as its directory value. When it reports the project behind — or any CLI output already shows it (the stderr notice This project pins hyperframes@… (latest …), or _meta.updateAvailable: true in a --json result from a pinned script) — apply with npx hyperframes@latest upgrade --project ., then verify with npx hyperframes check. A passing check confirms the project's compositions still validate on the new version — not that rendered output is frame-identical to the old pin — so a successful bump is never silent: name the old and new version in the run's summary. A project with no composition yet needs no verification. If the check fails, revert the package.json change, continue on the pinned version, and report which version the project stays on and why. Act on the signal rather than relaying it to the user; never leave a bumped pin unverified.

2. Route fresh creation

Use the first matching row. Match the requested deliverable, not a word or file type mentioned in passing.

Before finalizing the route, read references/routes/<workflow>.md — one small file per route: the canonical input/output/trigger contract (available before lazy-installed workflow skills are present) plus that route's interview entry. If the candidate does not satisfy its contract, continue routing instead of forcing the match. Read only the matched route's file.

Resolve common ambiguities

• A short animated title, logo sting, stat hit, chart hit, map hit, or standalone lower-third is /motion-graphics when it is unnarrated and motion is the message. A static title card, narrated sequence, longer montage, or custom loop is /general-video.
• An explicitly short motion graphic may use a URL, tweet, article, or screenshot as source material. A generic "make a video from this site" request is /product-launch-video.
• Existing footage with captions routes to /embedded-captions; footage with designed information cards routes to /talking-head-recut. Retiming, reordering, recoloring, reframing, or remixing footage is a custom edit and falls through to /general-video.
• A music file selects /music-to-video only when its beat grid drives the piece. Music used as a bed does not override the subject-matched route.
• "I want a storyboard" changes the review process, not the workflow. With no other routing signal, use /general-video. A confirmed sketched board may itself be the requested deliverable; the review loop defines that stop point.
• Specialized narrative workflows support up to about 3 minutes and are strongest around 30–90s. Route a clearly longer piece to /general-video. Length never overrides an explicit port, deck, caption, overlay, or music-driven deliverable.

3. Route once, then leave

For fresh creation the intent layer (references/intent-interview.md) runs the full conversation — memory, triage, pitch round, must-haves, run-shape, hand-off — and ends by writing BRIEF.md. The brief is the only routing artifact the workflow reads; nothing later re-opens this skill or the interview. Answer every later "what did the route require?" from BRIEF.md.

4. Install and enter the workflow

Before reading the selected workflow, install or refresh it and the core domain skills:

npx hyperframes skills update <workflow-name>

Use the bare name without /. If the command fails, surface the error; do not reconstruct the workflow from memory. Everything else about installation — the core-vs-lazy split, what init refreshes, diagnosis, CI opt-out, and the no-CLI fallback — lives in references/skill-lifecycle.md.

5. Load domain skills on demand

Creator edit phrases are cross-domain requests. Load every skill named in the matching row:

Constant data-playback-rate is render-safe for picture and pitch-preserved sound. It does not make source speed ramps keyframeable; preprocess ramps. For copyable edit contracts, load /hyperframes-core → references/creator-editing-recipes.md.

Broad feedback about how photographic media looks or behaves also routes to /media-use, even when the user never says “color grading” or “effect”: fix dark/flat/boring footage, stylize a clip, hide a face, or improve a media reveal. Read ../media-use/references/media-treatments.md before editing a treatment; it governs how footage is treated, never whether media may be used. Do not substitute a generic LUT, CSS filter/overlay, or opacity tween for an existing canonical treatment primitive. Keep text/layout/motion-only edits in their owning domain. During a build with important photographic media, include one grounded media-polish scan in the final quality pass; leaving suitable media unchanged is a valid result.

Domain skills never take ownership of the end-to-end deliverable. Load only what the active workflow needs.