Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Programmatic video creation in React with Remotion. Use this skill whenever the user writes Remotion, npx create-video, remotion studio/render/still, useCurrentFrame/interpolate/Sequence, Composition, Mediabunny, or any React-driven video/animation/caption/motion-graphics task — including scaffolding a project, writing markup, rendering/exporting (incl. transparent), adding captions/subtitles, making Studio-editable animations, building a Remotion SaaS (Player/Lambda/Vercel/Cloudflare), or getti
.claude/skills/fradser-remotion/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | -37% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -46% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -13% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -23% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -32% | 0% |
This skill is a router for all Remotion video-creation operations. It mirrors the upstream remotion-dev/skills repository. Based on user intent, read the corresponding sub-skill SKILL.md before executing.
Sub-skills are authoritative. The files under each sub-skill directory are upstream content kept verbatim — do not edit them; re-sync from upstream instead. The table below is the index, the sub-skill SKILL.md is the instruction.
| Sub-skill | Entry | Use When | |-----------|-------|----------| | New Project | remotion-create/remotion-create.md | Scaffold a new Remotion project (npx create-video), design a video, start the Studio preview | | React Markup | remotion-markup/remotion-markup.md | Writing Remotion React markup — the core: animations, media, sequences, timing, fonts, effects, maps, Lottie, DOM measurement, transitions, trimming | | Rendering | remotion-render/remotion-render.md | Rendering videos and stills (npx remotion render/still), transparent videos | | Captions | remotion-captions/remotion-captions.md | Transcribing, importing SRT, and displaying captions/subtitles | | Interactivity | remotion-interactivity/remotion-interactivity.md | Making animations editable in Remotion Studio Visual Mode | | SaaS / Apps | remotion-saas/remotion-saas.md | Building video apps — <Player>, rendering on Lambda/Vercel/Cloudflare/Node, client-side rendering, choosing a framework | | Mediabunny | mediabunny/mediabunny.md | Multimedia handling in the browser — get audio/video duration, get video dimensions |
<name>/<name>.md, denested from SKILL.md so sub-skills are not auto-discovered), then follow any references/-style .md files it points to (each sub-skill folder carries its own detail files).../remotion-markup/remotion-markup.md — these resolve within this remotion/ directory.Some upstream sub-skill files contain links like [Remotion Best Practices](../remotion-best-practices/remotion-best-practices.md). remotion-best-practices/ is a local shim (not an upstream sub-skill) that forwards to this top-level router — treat those links as "go back to the Remotion router" (i.e. this file). All other sibling cross-references (../remotion-markup/remotion-markup.md, ../remotion-captions/remotion-captions.md, etc.) resolve correctly.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 21,094 | 28,481 | +35% | 1 | 1 | 0% | 3,410 | 4,684 | +37% | 0 | 0 | — |
case-02 | fail→fail | 15,216 | 11,922 | -22% | 1 | 1 | 0% | 2,762 | 2,862 | +4% | 0 | 0 | — |
case-03 | pass→pass | 7,189 | 4,244 | -41% | 1 | 1 | 0% | 747 | 1,353 | +81% | 0 | 0 | — |
case-04 | fail→pass | 11,472 | 3,161 | -72% | 1 | 1 | 0% | 2,123 | 1,339 | -37% | 0 | 0 | — |
case-05 | fail→pass | 15,755 | 3,679 | -77% | 1 | 1 | 0% | 2,596 | 1,405 | -46% | 0 | 0 | — |
case-20 | pass→pass | 18,613 | 14,783 | -21% | 1 | 1 | 0% | 3,046 | 3,144 | +3% | 0 | 0 | — |
case-06 | fail→pass | 10,006 | 3,616 | -64% | 1 | 1 | 0% | 1,515 | 1,318 | -13% | 0 | 0 | — |
case-07 | fail→pass | 10,442 | 4,007 | -62% | 1 | 1 | 0% | 1,945 | 1,496 | -23% | 0 | 0 | — |
case-08 | fail→pass | 13,883 | 3,853 | -72% | 1 | 1 | 0% | 2,059 | 1,399 | -32% | 0 | 0 | — |
case-09 | fail→pass | 12,757 | 3,494 | -73% | 1 | 1 | 0% | 2,223 | 1,341 | -40% | 0 | 0 | — |
case-21 | pass→pass | 17,864 | 15,046 | -16% | 1 | 1 | 0% | 3,230 | 3,358 | +4% | 0 | 0 | — |
case-10 | fail→pass | 12,333 | 3,377 | -73% | 1 | 1 | 0% | 2,063 | 1,389 | -33% | 0 | 0 | — |
case-11 | fail→pass | 9,897 | 3,555 | -64% | 1 | 1 | 0% | 1,562 | 1,470 | -6% | 0 | 0 | — |
case-12 | pass→pass | 7,063 | 4,096 | -42% | 1 | 1 | 0% | 1,046 | 1,377 | +32% | 0 | 0 | — |
case-13 | fail→pass | 14,031 | 3,325 | -76% | 1 | 1 | 0% | 2,634 | 1,388 | -47% | 0 | 0 | — |
case-19 | fail→pass | 11,105 | 2,410 | -78% | 1 | 1 | 0% | 1,792 | 1,227 | -32% | 0 | 0 | — |
case-14 | fail→pass | 15,998 | 5,876 | -63% | 1 | 1 | 0% | 2,469 | 1,786 | -28% | 0 | 0 | — |
case-15 | fail→pass | 11,158 | 3,180 | -72% | 1 | 1 | 0% | 1,834 | 1,378 | -25% | 0 | 0 | — |
case-16 | pass→pass | 7,075 | 12,078 | +71% | 1 | 1 | 0% | 1,075 | 1,078 | +0% | 0 | 0 | — |
case-17 | pass→pass | 11,068 | 2,697 | -76% | 1 | 1 | 0% | 1,608 | 1,245 | -23% | 0 | 0 | — |
case-18 | fail→pass | 8,518 | 5,226 | -39% | 1 | 1 | 0% | 1,383 | 1,649 | +19% | 0 | 0 | — |
case-22 | pass→pass | 16,171 | 15,332 | -5% | 1 | 1 | 0% | 2,486 | 3,267 | +31% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +59 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.