* perf(engine): fast capture stays on for frames a shadow, blur or clip effect does not touch A single drop-shadow, blur, backdrop-filter or blend element, or one clip-path or mask tween, sent the whole render to screenshot capture, 2-3x slower than drawElement. The two init gates now name the frames the effect can show on. Only those frames capture by screenshot, in a second browser opened on the first of them; every other frame keeps drawElement. * fix(engine): scan effects in a throwaway session; hand risk frames over only where captured The effect scan seeked the render's own page to the end, leaving its timeline state ahead of the render. It now runs in a throwaway session with the render's own seek. A hidden effect element with a visible descendant counts as shown. Per-frame capture is opt-in per caller (deFrameFallback), so disk, segmented, chunk and parallel-worker sessions keep the whole-render fallback, and a route that cannot hand frames over drops drawElement up front. * fix(engine): a timeline callback keeps the whole-render fallback; trim comment share A tl.call or onComplete can add an effect's class to any element mid-render, so the elements found at the start are not the only ones it can reach. Such a timeline is unbounded and keeps the whole-render screenshot fallback. * style(engine): shorten the live-frame scan comment * perf(engine): after a timeline callback, not before, every frame may show the effect A callback can add an effect's class to any element, so only the frames from the first callback on lose their element bounds. Frames before it keep the scanned bounds, so a late callback no longer sends the whole render to screenshot capture. * perf(engine): open the second screenshot browser only when memory is available for it Per-frame capture needs a second browser (479 MB measured at 1080p). When the render would leave less than 1.5 GB available after it, the render keeps the whole-render screenshot fallback instead. getAvailableMemoryMb moves from the CLI to the engine's system-memory module, which already owns the host's memory readings; os.freemem() reads 367 MB on this 49 GB Mac, so the guard uses free, inactive, speculative and purgeable pages. * fix(engine): budget the whole 1 GB per-frame capture adds, not just the second browser The render's process tree goes from 1002 to 2001 MB at 1080p: the screenshot browser (479 MB) plus a drawElement browser that is itself about 500 MB larger than a screenshot one. The guard budgeted 500 MB. It now budgets the 1 GB measured, scaled by output pixels.
hyperframes
CLI for creating, previewing, and rendering HTML video compositions.
Install
npm install -g hyperframes
Or use directly with npx:
npx hyperframes <command>
Requirements: Node.js >= 22, FFmpeg
Commands
init
Scaffold a new Hyperframes project from a template:
npx hyperframes init my-video
cd my-video
preview
Start the live preview studio in your browser:
npx hyperframes preview
# Studio: http://localhost:3002/#project/my-video
# Server: http://localhost:3002
npx hyperframes preview --port 4567
In an interactive terminal, the preview stays attached until you press
Ctrl+C. In a non-interactive shell such as a coding-agent session, the same
command starts a managed preview that survives after the command returns. Use
--background or --foreground to choose explicitly, and manage persistent
previews with --status, --stop, --list, and --kill-all. Add --json to
managed lifecycle commands for machine-readable output. --foreground --json
prints the ready-session envelope once, then remains attached until stopped.
audio
Measure where sounds land: onsets, the moment each sound peaks, loudness and silence. Give files, or query the film's own clips in film time:
npx hyperframes audio sfx/whoosh.mp3
npx hyperframes audio --at 12 --json
normalize-audio
Measure two local authored audio clips with integrated LUFS and match the target
to the unchanged reference. The command is a dry run unless --write is passed:
npx hyperframes normalize-audio --reference target-audio --target user-audio
npx hyperframes normalize-audio --reference target-audio --target user-audio --write
It updates only the target element's data-volume and refuses unsafe boosts
that exceed Studio's +12 dB ceiling or would clip.
render
Render a composition to MP4. Run from the project directory; the positional
argument is the project directory (not a file), so render the project's
index.html directly, or point at a specific composition file with -c:
npx hyperframes render -o output.mp4
npx hyperframes render -c ./my-composition.html -o output.mp4
A render stops when the process that started it exits. To keep one running after
its launcher exits, set HYPERFRAMES_RENDER_DETACHED=1 and start it in its own
process group, e.g. bash -c 'set -m; HYPERFRAMES_RENDER_DETACHED=1 npx hyperframes render -o out.mp4 < /dev/null > render.log 2>&1 &'.
Piped, non-quiet renders emit a single @hf-diagnostic JSON line when an audio
failure reaches the CLI error handler. Version 1 has operation: "render" and
at most 16 audioFailures: allowlisted stage, reason, owner, retryable, and
optional subprocess exit code, signal, termination reason and OS code. It omits
messages, stderr, source paths/URLs, element IDs and command arguments. Consumers
must ignore unknown versions and fields. This is local CLI output; uploading it
requires the caller's existing consent and privacy checks. Quiet mode (including
batch JSON mode) emits no diagnostic line.
This covers audio preparation and mix failures, including the typed external interruption path. Probe failures retain their existing stage/reason but do not gain subprocess evidence here; browser/FX failures remain classified as internal. Desktop ingestion and live report delivery are separate changes.
publish
Upload a project directory and get a hosted URL that keeps working after the CLI exits. Published projects are private by default:
npx hyperframes publish
npx hyperframes publish ./my-video
npx hyperframes publish --public
npx hyperframes publish --yes
Signed-out publishing returns an authentication-required claim URL; opening it
lets someone sign in and claim the project. Sign in first with
npx hyperframes auth login to publish an owned project you can update. Use
--public to make the claimed project visible to anyone. --yes only skips the
confirmation prompt and does not change visibility.
Signed-in publishers can use --update <url-or-id> to target an existing project
or --space <space-id> to publish into a shared team space. If the requested
project is missing or inaccessible, publishing can create a new project instead;
check the printed URL and status.
See the publish reference for all options, including video proxy settings.
lint
Validate your Hyperframes HTML:
npx hyperframes lint ./my-composition
npx hyperframes lint ./my-composition --json # JSON output for CI/tooling
npx hyperframes lint ./my-composition --verbose # Include info-level findings
By default only errors and warnings are shown. Use --verbose to also display informational findings (e.g., external script dependency notices). Use --json for machine-readable output with errorCount, warningCount, infoCount, and a findings array.
compositions
List compositions found in the current project:
npx hyperframes compositions
benchmark
Run rendering benchmarks:
npx hyperframes benchmark ./my-composition.html
doctor
Check your environment for required dependencies (Chrome, FFmpeg, Node.js):
npx hyperframes doctor
browser
Manage the bundled Chrome/Chromium installation:
npx hyperframes browser
info
Print version and environment info:
npx hyperframes info
docs
Open the documentation in your browser:
npx hyperframes docs
upgrade
Check for updates and show upgrade instructions:
npx hyperframes upgrade
npx hyperframes upgrade --check --json # machine-readable for agents
Documentation
Full documentation: hyperframes.heygen.com/packages/cli
Related packages
@hyperframes/core— types, parsers, frame adapters@hyperframes/engine— rendering engine@hyperframes/producer— render pipeline@hyperframes/studio— composition editor UI