animai
Enables AI agents to search for and fetch pre-vetted, production-safe 3D and motion components (React Three Fiber & GSAP) for injection into Next.js apps via AST-safe edits.
README
AnimAI
Production-safe 3D & motion components your AI agent can install.
AI agents are great at UI — and notoriously bad at WebGL. Ask one for a Three.js hero and you'll usually get VRAM leaks, SSR crashes, and hydration bugs. AnimAI fixes that with a registry of pre-vetted React Three Fiber & GSAP components that are certified leak-free, SSR-safe, and performance-budgeted — then injected into your Next.js app with AST-safe edits (never regex) by a CLI or an MCP server your agent calls directly.
npx animai add hero-orbital-rig
Why vetted matters
Every component ships with a machine-generated vetting report, produced by
the harness in scripts/vet.mjs and enforced in CI:
| Check | Guarantee |
|---|---|
| Dispose audit | a useEffect cleanup disposes every geometry/material/texture — no VRAM leaks |
| SSR gate | WebGL is loaded via next/dynamic { ssr: false } — no server crashes |
| GSAP rule | animations use useGSAP() — timelines revert cleanly in Strict Mode |
| Canvas sizing | <Canvas> always gets an explicitly sized container |
Use it from your AI agent (MCP)
Add one line to your agent's MCP config (Cursor, Claude Code, Windsurf):
{
"mcpServers": {
"animai": { "command": "npx", "args": ["-y", "animai", "mcp"] }
}
}
Your agent gets two flat, deterministic tools:
search_motion_library— natural-language search over the registryfetch_motion_component— exact vetted source + dependencies + integration steps
Use it from the CLI
npx animai search "dark 3d hero" # find components
npx animai add hero-orbital-rig # write files + AST-patch your route
npx animai add hero-orbital-rig --dry-run
npx animai add hero-orbital-rig --target main
npx animai doctor hero-orbital-rig # check peer-dependency versions
npx animai update hero-orbital-rig # re-fetch latest (guards your local edits)
npx animai login <license-key> # unlock Pro components
add writes the component + SSR-safe loader into app/components/ and injects
the JSX into your route using @ast-grep/napi Find & Patch — imports and nodes
are inserted through the AST, so your source is never corrupted by string
replacement.
Free & Pro
The CLI, MCP server, and a growing set of components are free (MIT). Pro components are delivered through a licensed registry — the public npm package and CDN never contain Pro source. See docs/registry.md.
Repository layout
| Path | What it is |
|---|---|
src/ |
CLI + MCP server + registry client (the animai npm package) |
registry/ |
Component source of truth (manifest.json + code per component) |
scripts/ |
Registry/CDN build steps + the vetting harness |
web/ |
animai.dev — browse, docs, and a live MCP playground |
worker/ |
Cloudflare Worker serving the licensed registry (Polar) |
docs/ |
vetting and registry architecture |
examples/next-demo |
Minimal Next.js target app for trying add locally |
Develop
npm install
npm run build # build registry data + compile TypeScript
npm run vet # run the vetting harness, write reports
npm run vet:check # CI gate — non-zero exit on any critical failure
npm run demo:dry-run # preview an AST injection against examples/next-demo
npm run mcp # start the MCP server on stdio
License
MIT for the tooling and free components. Pro components are licensed per the terms at animai.dev.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。