Two independent post-release feedback reports of the same mechanism from
two different skills (both delegate to this one shared engine): audio.mjs
fired every line's Kokoro TTS + whisper-transcribe subprocess concurrently
via a bare Promise.all with no cap.
- One OOM'd 12/13 lines on a resource-constrained laptop (32GB total, ~7GB
free), requiring a manual patch to a sequential for-loop.
- The other saw 7/8 lines fail on first run, then pass on retry once the
model was cached — concurrent cold-start model loads overwhelming the
machine, not a real synthesis failure.
Kokoro/Whisper each load their own local model per subprocess, so firing
every line at once multiplies that cost by the line count. Extracted the
concurrency cap into lib/concurrency.mjs (audio.mjs is a script — it runs
CLI/exit side effects on import, so it can't be unit-tested directly;
the cap is small enough to pull out and test in isolation). Default 4,
overridable via HYPERFRAMES_TTS_CONCURRENCY, floored at 1 (matching one
report's own manual workaround).
hyperframes-media/scripts/audio.mjs is the single canonical engine per its
own header comment; product-launch-video, faceless-explainer, and
pr-to-video each carry a thin wrapper that spawns this file as a
subprocess (confirmed via their DEFAULT_ENGINE path), so this one fix
covers all four skills without touching the other three.
Tests: 4 new cases for mapWithConcurrency (order preserved regardless of
completion order, cap actually enforced, limit > item count doesn't hang,
empty input). Full skills test suite (514 tests) shows no new failures —
the 444 pre-existing failures are environment-dependent and reproduce
identically on unmodified main.
* feat(hyperframes-creative): add frame-preset library
Add a library of ready-made visual frame presets (claude, biennale-yellow,
blockframe, blue-professional, bold-poster, broadside, capsule, cartesian,
cobalt-grid, coral, creative-mode, daisy-days, editorial-forest, …), each with
a FRAME.md spec, a frame-showcase.html, and a per-preset caption-skin.html.
Registered in the creative design-spec so workflows can remix a preset onto
brand tokens.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* feat(hyperframes-media): shared TTS/BGM/SFX audio engine
Add a shared audio engine under hyperframes-media (scripts/audio.mjs + lib/
tts.mjs, bgm.mjs, sfx.mjs, heygen.mjs) plus a bundled SFX pack and manifest.
Workflows resolve this engine by path (../../hyperframes-media/scripts/
audio.mjs) for text-to-speech, background music, and sound effects, so audio
is authored once and reused across skills instead of duplicated per workflow.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* feat(skills): gate render on user review; refresh router, core, general-video
- hyperframes-cli: render is now user-gated — preview opens Studio (the timeline
editor where the user can hand-edit anything, not just watch); never
auto-render once checks pass, pause at preview and render only after approval.
- hyperframes (router): tighten the entry SKILL.md description + routing.
- hyperframes-core: rewrite SKILL.md and add script-format.md + storyboard-format.md
references for the script-driven authoring architecture.
- general-video: tidy the fallback-workflow description and routing table.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* style(hyperframes-creative): reformat frame-preset showcase HTML
Run the HTML formatter over the frame-showcase.html files (indentation,
self-closing void tags, one CSS declaration per line). Formatting only — no
content or markup changes.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* fix(hyperframes-media): correct wait-bgm field mapping and guard credential parse
Two correctness fixes from review (#1632):
- wait-bgm.mjs read audioMeta.bgm_path / audioMeta.bgm_enabled, but audio.mjs
writes the path nested as bgm.path and the flag as bgm_pending. The detached
generate path (Lyria/MusicGen) therefore always saw an empty path and exited
status: disabled, silently dropping the music track even while generation was
running. Read audioMeta.bgm?.path and gate on bgm_pending.
- heygenCredential() had an unguarded JSON.parse despite documenting that it
never throws — a malformed ~/.heygen credentials file crashed the engine at
startup instead of degrading to no-credential. Wrap the parse and return null.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* chore(hyperframes): add router tag to entry skill metadata
Fold the router metadata tag into the foundation rewrite of the entry SKILL.md.
This file is owned by this PR (the full router rewrite); keeping the tag tweak
here — instead of a separate edit on the pre-rewrite version in another PR —
avoids a guaranteed merge conflict between the two.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Miguel Ángel