More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+22 more

    Video Craft — narrated video pipeline for HyperFrames

    2

    Turn one scene manifest into a narrated 16:9 or 9:16 video: voice-driven timings, motion, chapter wipes, SFX, music bed, a QA gate and a constant-bitrate render.

    $5

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned

    About this skill

    Video Craft is the reusable part of a real production: a 31-scene, 10-minute narrated explainer that shipped in three rendered cuts. It teaches your agent to build a video as a generated HyperFrames project, so a re-recorded line, a reordered scene or a vertical cut is a regenerate instead of a hand edit.

    How it works

    • One manifest (build/rows.js) defines every scene. A generator writes every composition, the root timeline, YouTube chapters, a VO timing CSV for a voice actor, and timeline.json.
    • Timings come from the narration: scratch TTS is measured, each scene stretches to fit its line, and reveals land on a phrase (cue(row, "phrase")), so a new voice moves every cue.
    • Motion vocabulary: entrances, a subtle drift on every scene, chapter wipe bands, pop/fade/count-up reveals and push-in camera moves in the generator, recipes for typing, word highlight and region glow, plus the lint footguns that cost re-runs.
    • Sound: five deterministic synthesised SFX (no licences), a synthesised lo-fi music bed sized to the film, a single automation lane for ducking and the end fade.
    • A QA gate script that prints PASS/FAIL: scene timing, media, typography, every narrated row voiced and no take cut short, the music lane silent where it should be, and each render checked against a fingerprint of the sources it was made from, so a stale render fails.
    • A detached render script for three deliverables (voice, music-only, vertical) with a constant-bitrate re-encode; it reports the number of failed steps.

    Requirements

    Node 22+, ffmpeg/ffprobe, the HyperFrames CLI (open source, npx hyperframes) and its agent skills. Scratch voice uses macOS say (any installed voice); on other systems put your own WAV files in first and the script only measures them. The template is Thai-first (Thai fonts and sample rows) but the pipeline is language-agnostic.

    Limits, stated plainly

    • Scratch TTS is for timing and pacing review, not for publishing.
    • Fonts are not bundled; add your own font files for an embedded render.
    • The vertical layout renders from the template, but no real vertical film has been reviewed end to end.

    How to use

    Skills on Agensi run inside your AI agent. Copy the prompt, paste it into your agent, and it connects to Agensi and installs this skill for you. If a payment is needed, your agent shows it to you first.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill passes our trust checks below.

    Security scanned

    Passed our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scanned

    Verified clean 5 days ago

    • Passed all security checks, Safe to install

    Requires

    hyperframes
    Node.js
    macOS

    Listed5 days ago

    What's inside

    Frequently Asked Questions