anime-cel
Skill18 starsChanged 49 days agoMITView on GitHub
Make Japanese TV-anime style videos (MP4) with code — cel-shaded characters with tapered line art, big detailed eyes, hair clumps with a specular band, comedic anime reactions (sweat drop, anger vein, sparkle eyes, ≥< eyes, dot eyes, tears), painted skies with cumulus clouds, school/street/classroom/rooftop sets with time-of-day palettes, speed/focus lines, impact frames, dramatic zooms, smears, sakura petals and anime subtitles (CJK). Canvas 2D rendered frame by frame in headless Chrome, encoded with ffmpeg. Use when a reference video or brief calls for anime / 動畫 / 日系動畫 / アニメ / cel-shaded / slice-of-life / school romance / shōnen-style reactions, or when the video-clone router picks the anime-cel style. 2D only; characters are original (Hina, Ren) or new ones built on the same rig.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- edenfunf/reelmimicword for word, first seen Nov 4, 2026
Other versions1 project uses a version of this.
Word for word · 1
- edenfunf/reelmimic★ 2k100% same
Edited · 0
None found.
Wrote their own version · 0
None found.
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository37 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- blender-product-filmSkill
- captions-overlaySkill
- crayon-storybookSkill
- cut-the-curveSkill
- embedded-captionsSkill
- faceless-explainerSkill
- ffmpegSkill
- figmaSkill
- frontend-designSkill
- general-videoSkill
- hyperframes-animationSkill
- hyperframes-audioSkill
- hyperframes-cliSkill
- hyperframes-coreSkill
- hyperframes-creativeSkill
- hyperframes-keyframesSkill
- hyperframes-registrySkill
- hyperframesSkill
- hyperframes-studioSkill
- media-useSkill
- motion-doctrineSkill
- motion-graphicsSkill
- music-to-videoSkill
- oversized-cursorSkill
- painted-animationSkill
- paper-cutoutSkill
- pixel-artSkill
- product-launch-videoSkill
- pr-to-videoSkill
- remotion-to-hyperframesSkill
- seam-craftSkill
- slideshowSkill
- talking-head-recutSkill
- video-cloneSkill
- whiteboardSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 0 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 12
Read GUIDE.md in full before building anything.
- Very specific: line 64
Captions: only via overlayHook + subtitle() (white, dark outline, CJK font).
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
anime-cel
Modern TV-anime look drawn entirely by code: characters with confident tapered line art (heavier on the shadow side),
flat cel colour + one hard-edged shadow tone + a rim highlight, large layered eyes, hair built from pointed locks with
an "angel ring" highlight; line-free painted backgrounds; anime staging effects. Every frame is a pure function of t.
Read GUIDE.md in full before building anything. Model sheets in docs/: faces-hina.jpg,
faces-ren.jpg (19 expressions each), turns.jpg (front → 3/4 → profile), bodies.jpg, motion.jpg (stances, walk, run), hands.jpg, sets.jpg,
fx.jpg, and demo-sheet.jpg (the demo film).
Workflow
- Scaffold
bash .claude/skills/anime-cel/scripts/new_project.sh <project-dir>(--keep-demokeeps the example). It copiestemplate/, unhooks the demo, runsnpm install, checks node/ffmpeg/Chrome. - Storyboard each shot with its reads (what the viewer must understand, in order, with times) and the anime
staging you will use (establishing pan, cut on action, focus-line BG for the emotional beat, impact frame, reaction
close-up…). Set
durationinsrc/config.js. - Build one file per shot in
src/scenes/(IIFE ending inshots([[t0, fn]]), add its<script>tostudio.html). A shot paints the whole frame:setTOD()→SET.<set>(cam, t)(opens the camera) → characters →camEnd()→ screen effects → transition. Captions go inwindow.overlayHookwithsubtitle(). - Look at it (from the project dir), open every image with Read:
node render.mjs --sheet=0.5,1.5,2.5,3.5 --cols=4 --w=480 --out=out/check/a.jpg # key frames node render.mjs --strip=3.8:4.5 --cols=6 --w=320 --out=out/check/s.jpg # every frame of an action node render.mjs --stills=4.4 --out=out/check/st # full-res PNG to judge faces node render.mjs --loop=faces_hina --sheet=0 --cols=1 --w=1920 --out=out/check/f.jpg # model sheets (faces_ren, turns, bodies, motion, hands, sets, fx, close_hina, close_ren) - Render
node render.mjs --frames --workers=4 && node render.mjs --encode --out=out/video.mp4 [--audio=...].
API summary
hina(x, y, u, o) / ren(x, y, u, o) → { head, handL, handR, footL, footR, top, u } // (x, y) ground point, u = head height px
bust(DESIGNS.ren, x, y, u, o) // close-up: (x, y) = head centre
o = { t, turn (-1..1: 0 front, ±.5 3/4, ±1 profile), headTurn, bodyTurn, face ('happy' …), look [x,y], blink, mouth, talk,
armL/armR [shoulder°, elbow°, hand, wrist°, foreshorten], legL/legR [hip°, knee°], walk: phase, run: phase,
lean°, bend°, tilt°, sq, dy, flip, sway [x,y], fx ['sweat','vein','steam','gloom','sparkles'], blush,
prop { hand:'L'|'R', draw: PROPS.envelope, rot°, s, at }, armsOver, ground:false,
stance:'contra'|'straight', weight:'L'|'R', contra } // hands: open relax fist point peace spread hold grip pocket
faceAt(t, [[0,'neutral'],[1.2,'surprised']]) · autoBlink(t, seed) · talk(t) · follow(t, f) · walkPose/runPose · stroll
setTOD('day'|'morning'|'evening'|'night') · setLight(x, y) · SET.street / SET.classroom / SET.rooftop (cam, t, o)
focusLines · speedLines · impactFrame · flash · dramaticZoom · smear · smearLines · blurBand · sparkle(s) · petals
slashWipe · irisWipe · subtitle(c, txt, o) · onTwos(t) · kit: kf, seg, ease*, spring, jump, take, camBegin/camEnd
drawing primitives: ink (tapered line), tubePts, ribbonPts, crPts, cel, drawGroup, layer (cache), paintPolyRules that matter most
- Pure function of t — no
Math.random(), no state between frames (rnd(key)/hash();follow()for springs). - Scenes never draw body parts. Characters are one call. New characters = a new entry in
DESIGNS+HAIR(+ clothes branch) incast.js, never ad-hoc drawing in a scene. - Animate characters on twos (
const ta = onTwos(lt)for poses/faces), keep camera moves smooth (lt). - Anime timing: quick actions with a held pose after; anticipation → snap → overshoot (
spring,take); faces change on a drawing viafaceAt(it adds the blink + pop), never by crossfading. - Staging: establish → react. Emotional peaks cut the BG to a flat gradient +
focusLines; hits get animpactFrame(2–3 frames) +dramaticZoom; fast moves getsmearLines. - Sizes: full body u≈110–150, knees-up 180–260, bust 380–560. Keep faces ≥ u 150 when they carry a read.
- Transitions at every seam (
slashWipe,irisWipestar,flash, impact cut, cut on action). - Captions: only via
overlayHook+subtitle()(white, dark outline, CJK font). Push new caption text intowindow.EXTRA_FONTSso the glyphs load before rendering.
Engine gotchas
SET.*leaves the camera open; draw characters in the same world coords and callcamEnd()before screen effects. Screen-space effects (blurBand,impactFrame,slashWipe) usesetTransformand ignore the camera.- Characters face screen-right with positive
turn. With|turn| > .3limb angles are "forward/back" in the facing direction; in front view they are "outward/inward". The near arm when facing right isarmL. - Props in hands: use hand
'hold'(pinch, prop extends past the fingertips) or'grip'(handle in the fist).rotrotates the prop about the grip point. layer(key, …)caches static paintings per page; include the time-of-day name in the key if colours depend on it.- Speed: ~5–25 ms to draw a frame, ~100 ms/frame effective incl. JPEG with 4 workers (first frame ~0.6 s builds caches).
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

