treemation
Agent skill: turn a topic into a narrated 9:16 3D educational video (TikTok/Reels/Shorts) — script, pixel-skin host, lip sync, captions, thumbnail, MP4
Safety scan
A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.
Badge for your README
Markdown
[](/listings/adytm404--treemation)HTML
<a href="/listings/adytm404--treemation"><img src="/api/listings/adytm404--treemation/badge.svg" alt="findsafeskills safety rating" /></a>Install
Install instructions from the README
As written by the repo's author. findsafeskills has not verified these steps.
Claude Code (plugin)
Run this inside Claude Code (version 2.1.275 or later):
/plugin install treemation --marketplace Adytm404/treemation
Or in two steps, which also works on older versions:
/plugin marketplace add Adytm404/treemation /plugin install treemation@treemation
From a terminal instead:
claude plugin marketplace add Adytm404/treemation claude plugin install treemation@treemation
Update later with `/plugin marketplace update treemation`.
Any agent (skills.sh)
npx skills add Adytm404/treemation --skill treemation
> Run the command above, select your AI agent when prompted, then restart the agent so it can detect the skill.
What was scanned
The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.
.claude-plugin/marketplace.json · Manifest · 27 lines
- {
- "name": "treemation",
- "owner": {
- "name": "Adytm404",
- "url": "https://github.com/Adytm404"
- },
- "metadata": {
- "description": "Treemation: 3D educational video generator skill for Claude Code",
- "version": "1.3.0"
- },
- "plugins": [
- {
- "name": "treemation",
- "source": "./",
- "description": "Turn a topic into a narrated 9:16 3D educational video (TikTok/Reels/Shorts) with script, pixel-skin host, lip sync, captions, thumbnail, and MP4 render.",
- "version": "1.3.0",
- "category": "productivity",
- "keywords": [
- "video",
- "animation",
- "3d",
- "tiktok",
- "elevenlabs"
- ]
- }
- ]
- }
README.md · README · 172 lines
- # treemation
- 📦 [View on skills.sh](https://skills.sh/Adytm404/treemation/treemation)
- **Bikin video edukasi 3D vertikal (TikTok / Reels / Shorts) sepenuhnya dengan code — dari satu kalimat topik sampai MP4 + thumbnail.**
- An AI agent skill that turns a topic into a narrated, animated 9:16 explainer video: script for ElevenLabs, scene plan, a 3D stage with a pixel-skin host and mascot, word-by-word captions, vowel lip sync, sound effects, adaptive backgrounds, and a cover thumbnail. Built for Indonesian-language educational content.
- ## Demo
- All clips below were rendered by this skill (no video editor, no stock footage).
- <table>
- <tr>
- <td align="center"><img src="docs/demo/hook.gif" width="180" alt="Hook: close-up of the host with chips appearing"><br><sub>Hook 0–3 s</sub></td>
- <td align="center"><img src="docs/demo/datacenter.gif" width="180" alt="Data center building with server racks popping in"><br><sub>3D props</sub></td>
- <td align="center"><img src="docs/demo/gpu.gif" width="180" alt="GPU rig with spinning fans"><br><sub>Custom props</sub></td>
- <td align="center"><img src="docs/demo/website.gif" width="180" alt="Files flying into a server rack"><br><sub>UI + 3D sync</sub></td>
- <td align="center"><img src="docs/demo/spam.gif" width="180" alt="Spam pages swarming a server"><br><sub>Adaptive background</sub></td>
- </tr>
- </table>
- ▶️ **With sound:** [docs/demo/sample-datacenter.mp4](docs/demo/sample-datacenter.mp4) — a 15-second scene from *"Gimana data center bekerja?"* (narration generated with ElevenLabs).
- <table>
- <tr>
- <td align="center"><img src="docs/demo/thumb-datacenter.jpg" width="220" alt="Thumbnail: Gimana data center bekerja?"></td>
- <td align="center"><img src="docs/demo/thumb-judol.jpg" width="220" alt="Thumbnail: Kok ada judol di website kampus?"></td>
- </tr>
- <tr>
- <td align="center"><sub>Auto-generated cover</sub></td>
- <td align="center"><sub>Safe-area checked for profile grids</sub></td>
- </tr>
- </table>
- **3D sets** — when the script talks about a place, the whole room is built behind the host and the camera flies inside:
- <img src="docs/demo/set-datacenter.png" alt="Data center set: server aisle, then the camera moves into the cooling unit" width="480">
- **Character presets** (original pixel-skin designs, fully customizable — hair, hijab, outfit, colors, accessories):
- <img src="docs/demo/characters.png" alt="Eight pixel-skin character presets" width="720">
- ## Install
- ### Claude Code (plugin)
- Run this inside Claude Code (version 2.1.275 or later):
- ```
- /plugin install treemation --marketplace Adytm404/treemation
- ```
- Or in two steps, which also works on older versions:
- ```
- /plugin marketplace add Adytm404/treemation
- /plugin install treemation@treemation
- ```
- From a terminal instead:
- ```bash
- claude plugin marketplace add Adytm404/treemation
- claude plugin install treemation@treemation
- ```
- Update later with `/plugin marketplace update treemation`.
- ### Any agent (skills.sh)
- ```bash
- npx skills add Adytm404/treemation --skill treemation
- ```
- > Run the command above, select your AI agent when prompted, then restart the agent so it can detect the skill.
- ## How it works
- The skill runs in two steps. Each step is triggered by a single message.
- ### 1. Topic → script
- ```
- buat konten edukasi: apa itu domain, CTA ke tokoku.id
- ```
- The agent replies directly with:
- - A ready-to-paste **ElevenLabs script** (numbers and acronyms written the way they are spoken, e.g. "zenhosta dot com" → caption shows `zenhosta.com`).
- - **Three hook options** of different types (curious question, misconception, "what if", analogy, …) checked against hook rules for the first 3 seconds.
- - The **host character** chosen from your context (or described by you), plus mascot.
- - A **thumbnail** plan and a **scene table** with timing, camera shots, background theme, and visuals.
- - `video-plan.json` with every scene, shot, gesture, sound effect, and per-word caption timing.
- ### 2. MP3 → video
- Generate the narration in ElevenLabs, then send the MP3. The agent processes it end to end without further questions:
- 1. **Aligns** the plan to the audio (silence detection + punctuation, no AI model needed).
- 2. Writes `scenes.js` — 3D props on a floating stage and UI panels timed to the exact word that explains them.
- 3. Renders **preview stills**, checks them, and fixes overlaps.
- 4. Renders all frames with headless Chromium, builds the **thumbnail**, mixes **sound effects** under the narration, and encodes a **1080×1920 30 fps H.264 MP4**.
- ## Features
- | | |
- |---|---|
- | 🎬 Dynamic camera | Close-up, medium, mascot close-up, illustration, wide, low angle — cuts and moves on the right word |
- | 🧍 Pixel-skin host | Blocky game-style character with pixel anime face; 8 presets, hijab, outfits, skin tones, accessories |
- | 👄 Vowel lip sync | Mouth shapes A · I · U · E · O follow each word's vowels and close on pauses; foreign words read by pronunciation |
- | 🎥 Clear camera | Characters vanish or slide straight out of frame before the camera focuses on props — never walking across the shot |
- | 🎨 Adaptive background | Theme, particles (bokeh, stars, bubbles, snow, binary code) and motif icons change per scene |
- | 🏙️ 3D sets | Whole places (data center, office, warehouse, city, or your own) built behind the host; the camera flies to the spot being mentioned |
- | 🪝 Hook rules | 3-layer hook (visual, voice, on-screen text) with an automatic lint before the script is sent |
- | 💬 Captions | Big word-by-word karaoke captions, keywords highlighted in yellow, zoom punch on key words |
- | 🎯 Precise timing | Optional Whisper word timestamps (`--whisper`) for exact caption and lip-sync timing |
- | 🎭 Acting | Expressive poses (explain, clap, hands on hips), mascot reactions, squash-and-stretch, cartoon outlines |
- | 🔊 Sound effects | Whoosh, pop, ding, chime, typing, count-up, boom… generated procedurally and mixed under the voice |
- | 🎵 Background music | Your own track or a royalty-free code-generated one (`--music auto`), ducked under the narration |
- | ⚡ GPU rendering | Optional `--gpu` render (3–10× faster on machines with a graphics card) |
- | 🖼️ Thumbnail | 1080×1920 cover with title, props, frozen pose, and a safe-area check for 3:4 / 4:5 / 1:1 crops |
- ## Prerequisites
- - **Node.js + npm** — installs three.js, fonts, and icons per project
- - **Python 3** with `numpy` (and `pillow` for the thumbnail safe-area check)
- - **ffmpeg**
- - **Playwright Chromium** — `pip install playwright && python -m playwright install chromium`
- - *Optional:* `faster-whisper` for precise word timing (`pip install faster-whisper`, first model download ±500 MB)
- The setup script checks these and tells the agent what is missing.
- ## Supported agents
- Any agent that supports the `SKILL.md` format, including Claude Code, Codex, Cursor, and OpenCode.
- ## Project layout
- One video = one folder:
- ```
- video-apa-itu-domain/
- ├── plan.json # script, scenes, shots, characters, hook, thumbnail
- ├── naskah-elevenlabs.txt # text to paste into ElevenLabs
- ├── video-plan.json # timing (estimated, then aligned to the audio)
- ├── narasi.mp3 # your ElevenLabs narration
- ├── scenes.js # props + UI for this video
- ├── thumbnail.png
- └── <name>.mp4
- ```
- ## Skill contents
- ```
- skills/treemation/
- ├── SKILL.md
- ├── references/ # script & plan schema, hook research, characters, engine API, reply format
- ├── scripts/ # setup, timing, align, render, thumbnail, encode, character preview, project finder
- ├── assets/engine/ # three.js engine, pixel-skin characters, 3D set library, page template
- └── examples/ # a complete 69-second example (plan.json + scenes.js)
- ```
- ## Notes
- - Scripts, captions, and replies are in **Bahasa Indonesia**.
- - Characters are original designs. The skill avoids copying characters, skins, or mascots from other creators, games, or brands.
- - Educational content should be accurate: the skill writes example numbers from documentation ranges or clearly labeled data, and CTAs only state what you provide.
- ## License
- [MIT](LICENSE) © 2026 Adytm404
Ratings and comments
Sign in to rate this skill.
Sign inFree to use. If it’s helped, pass it on.