genable-mcp
Write-side MCP server for Figma — build, edit, and search Figma designs from Claude Code, Cursor, Cline, or any MCP client. Complements Figma's official read-only MCP with 41 tools for tree creation, variables, components, and visual verification.
README
<div align="center">
<img src="./assets/logo.svg" alt="Genable" width="96" />
Genable
Hand off the design busywork. Detailed prompt in. Fully editable Figma layers out — components typed, variables bound, variants generated, pages laid out.
<img src="./assets/cover.png" alt="Skip the busywork. Keep the design." width="100%" />
</div>
Why Genable
Most AI design tools ship a flat image or a templated mock. Genable is agentic — it plans, executes against Figma's scene graph through typed tools, and verifies its work. The output is real, editable Figma layers, not screenshots.
- Real layers — Frames with Auto Layout, real Text, real Components — not flattened images.
- Variables, not pasted hex — Color, typography, and spacing bound to variables with light and dark modes.
- Variants together — Light, dark, and brand-color themes generated in one pass, switchable in a click.
- Whole pages — Hero, pricing, features, FAQ laid out as proper sections — not a slab of nodes.
- Detailed prompts welcome — We don't sell prompt brevity. The more specific you are, the better the result.
What you can hand off
<img src="./assets/screenshot-1.png" alt="Six things Genable takes off your plate" width="100%" />
Three steps
<img src="./assets/screenshot-2.png" alt="From describing to designing — three steps" width="100%" />
Two surfaces, one engine
Use Genable as a Figma plugin (designers) or as the genable-mcp MCP server (Claude Code, Cursor, Cline, Continue, Zed, or any MCP-compatible agent). Same engine, same 41 tools.
Install the plugin
Install from Figma Community →
- Open Figma.
- Run
Plugins → Genable. - Paste an API key in Settings (any of the protocols below).
- Type a detailed prompt. Hit generate.
Install the MCP server
For agent-driven workflows. Pair with the official Figma MCP for full read + write.
{
"mcpServers": {
"genable": { "command": "npx", "args": ["-y", "genable-mcp"] }
}
}
41 tools — JSX tree creation, variables, components, cross-page navigation, visual verification. Full docs on npm →
Bring your own model
Genable speaks three protocols natively. Pick whichever you have keys for — keys stay on your device.
| Protocol | Examples | Get a key |
|---|---|---|
| Google Gemini | Gemini 2.5 Pro / Flash | aistudio.google.com |
| Anthropic Claude | Claude 4.7 Sonnet / Opus | console.anthropic.com |
| OpenAI-compatible | OpenRouter, DashScope (Qwen), Kimi K2.5, custom endpoints | varies |
Switch models any time from the Settings panel.
Examples
Detail helps. Try prompts like:
- "SaaS pricing page — 3 tiers, monthly/annual toggle, featured Pro plan, dark mode, brand color bound to variables."
- "Analytics dashboard — sidebar, KPI grid with sparklines, a hot-state table row, dark + light variants, bound to brand tokens."
- "Mobile onboarding — 3 screens with progress dots, illustrations, brand color blue (#4F90EE)."
- "Landing page hero — headline + subline + dual CTA + code preview card, Inter typography, 8pt spacing scale."
Genable returns a real, editable Figma frame you can drop straight into a design system.
Sponsor
Genable is built and maintained by one developer, in the open. If it saves you time or replaces a paid tool, please consider sponsoring:
Sponsorship pays for development time, model API quotas during testing, and ongoing improvements.
License
MIT — free for personal and commercial use.
<div align="center"> <sub> <a href="https://www.figma.com/community/plugin/1583731690321161934/genable-ai-ui-design-generator-prompt-to-ui-dashboard-landing-page-mobile-app">Figma plugin</a> · <a href="https://www.npmjs.com/package/genable-mcp">MCP server</a> · <a href="https://www.patreon.com/c/musec">Sponsor</a> </sub> </div>
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。