このSkillで書き出した映像
Kinetic Type Manifesto v1.1.0
言葉がビートに合わせて切り替わり、滑り、着地するタイポグラフィのブランド映像。
- フォーマット
- 動画 · 9:16
- 尺
- 0:15
マニフェストを書く、または貼り付けるだけで、キネティックタイポグラフィ映像に。フレーズごとに変わるレイアウト、規律ある書体システム、アクセントカラー、音楽に同期したトランジション。素材は不要、言葉と視点だけで作れます。
2026年10月25日 更新
得られるもの
- H.264 MP4 films in 9:16, 16:9, 1:1 or 4:5 from one script
- A poster frame for each film
- An editable scenes.json you can re-render any time
- Validation for reading speed, length and colour contrast
向いている用途
- Launch teasers and brand manifestos without footage
- Short vertical ads for TikTok, Reels and Shorts
- The same message localized for several markets
向いていない用途
- Product demos that need real screen recordings
- Talking-head or voice-over-led videos
- Ads that rely on stock or filmed footage
このSkillについて
/kinetic-type-manifesto Make a 15-second vertical teaser for our app launch in English and Japanese, black and orange, ending on “Start free”.
Typographic motion films from words alone. Each line is a shot: words rise from masks on a two-typeface system, one accent colour carries emphasis, and colour wipes hide every cut. The renderer is a pure function of time, so frames are exact and repeatable.
How it works
- Brief — your agent collects the message, audience, languages, formats, brand colours and CTA, asking only for what is missing.
- Script — it writes the lines first (one idea per scene) and waits for your approval. Other languages are adapted, not translated word for word.
- scenes.json — the approved script becomes a small JSON file with six scene types: statement, words, list, stat, quote and cta.
- Validate — a checker flags reading speed that is too fast, overlong lines, missing CTA and low colour contrast.
- Draft — a half-resolution render and a poster frame in seconds, for review.
- Final — H.264 MP4 at full resolution, re-targeted to 9:16, 16:9, 1:1 or 4:5 from the same script, with optional licensed music.
What makes it production-ready
- Text auto-fits the frame and stays out of TikTok/Reels/Shorts UI on vertical video.
- CJK languages use weight instead of fake italics; line breaking follows each language.
- Numbers in stat scenes count up; logos and URLs land on the end card.
- No API keys and no paid services: everything renders on your machine.
Tips
- Keep statements under eight words; split long ideas into two scenes.
- One emphasis phrase per scene reads better than three.
- Render a draft for each language before the finals.
始める前に必要なもの
- 用意するもの
- Your message and audience, the languages and formats you need, brand colours and name, a CTA, and optionally a logo and a music track you are licensed to use.
- 依存関係
- Node.js 20 or newer — required · Google Chrome (or CHROME_PATH) — required · FFmpeg — bundled via ffmpeg-static after npm install
- 外部コスト
- None. Everything renders on your machine; no API keys or paid services are used.
- 最終テスト
- v1.1.0 — Claude Code / Codex / Cursor
ユーザーレビュー
まだレビューはありません。インストールして最初に評価しましょう。
よくある質問
Do I need FFmpeg installed?
No. Running npm install in the Skill folder adds ffmpeg-static, a bundled FFmpeg binary. You only need Node.js 20+ and Google Chrome.
Can I use my brand fonts?
Yes. Name any Google Font or installed font in scenes.json; the renderer falls back to system fonts, including CJK, if a font is unavailable.
How long does a render take?
A 15-second 1080×1920 film takes roughly 45–60 seconds on a laptop; draft renders take a third of that.
Which languages work?
Any language Chrome can render. The samples on this page were rendered in English, Spanish, Japanese, German, Portuguese and Chinese from the same template.