ramensv/figma-mcp-go

浏览器自动化

本页由 AI 翻译，技术细节请以原文为准。查看原文

44 stars 243 forks 质量 53 趋势 13

Repository Install

ramensv/figma-mcp-go 是从 https://github.com/ramensv/figma-mcp-go. 导入的 MCP 服务器候选

概览

ramensv/figma-mcp-go 是从 https://github.com/ramensv/figma-mcp-go. 导入的 MCP 服务器候选

概述（AI 翻译）

一个零成本、无限的Figma MCP服务器——没有API密钥，没有速率限制，只有一个Go二进制。

支持光标、Claude、GitHub Copilot 以及任何兼容 MCP 的 AI 工具。

为什么会有这样的存在

大多数Figma MCP服务器依赖于Figma REST API。

听起来不错…直到你遇到这个：

计划	极限
启动 / 查看 / 协作	每月6次工具调用
Pro / Org（开发席位）	每天200次工具调用
企业号	每天600次工具调用

如果你在尝试AI工具，几分钟内就能消耗掉。

我没钱支付更高的限额。所以我做了一个完全不用API的东西。

布置

观看视频：

1.配置你的AI工具

.vscode/mcp.json

.mcp.json

2.安装Figma插件

在 Figma Desktop 中：插件→开发 →从 manifest 导入插件
在plugin.zip中选择manifest.json
在任何Figma文件中运行插件

可用工具

文件与选拔

工具	描述
`get_document`	完整的当前页面树
`get_metadata`	文件名，页面，当前页面
`get_selection`	当前选择的节点

本页由 AI 翻译，技术细节请以原文为准。

README

figma-mcp-go

A zero-cost, unlimited Figma MCP server — no API key, no rate limits, just a single Go binary.

Works with Cursor, Claude, GitHub Copilot, and any MCP-compatible AI tool.

Why this exists

Most Figma MCP servers rely on the Figma REST API.

That sounds fine… until you hit this:

Plan	Limit
Starter / View / Collab	6 tool calls/month
Pro / Org (Dev seat)	200 tool calls/day
Enterprise	600 tool calls/day

If you’re experimenting with AI tools, you’ll burn through that in minutes.

I didn’t have enough money to pay for higher limits. So I built something that doesn’t use the API at all.

Setup

Watch the video:

1. Configure your AI tool

.vscode/mcp.json

{
	"servers": {
		"figma-mcp-go": {
			"type": "stdio",
			"command": "npx",
			"args": [
				"-y",
				"@vkhanhqui/figma-mcp-go"
			]
		}
	}
}

.mcp.json

{
  "mcpServers": {
    "figma-mcp-go": {
      "command": "npx",
      "args": ["-y", "@vkhanhqui/figma-mcp-go"]
    }
  }
}

2. Install the Figma plugin

In Figma Desktop: Plugins → Development → Import plugin from manifest
Select manifest.json in the plugin.zip
Run the plugin inside any Figma file

Available tools

Document & Selection

Tool	Description
`get_document`	Full current page tree
`get_metadata`	File name, pages, current page
`get_selection`	Currently selected nodes
`get_node`	Single node by ID
`get_nodes_info`	Multiple nodes by ID
`get_design_context`	Depth-limited tree — token-efficient for large files
`scan_text_nodes`	All text nodes in a subtree
`scan_nodes_by_types`	Nodes matching given type list

Styles & Variables

Tool	Description
`get_styles`	Paint, text, effect, and grid styles
`get_variable_defs`	Variable collections and values
`get_local_components`	All components in the file
`get_annotations`	Dev-mode annotations

Export

Tool	Description
`get_screenshot`	Base64 image export of any node
`save_screenshots`	Export images to disk (server-side, no API call)

Contributing

Issues and PRs are welcome.

References

在 GitHub 查看完整 README

安装

npx -y @vkhanhqui/figma-mcp-go

新手？先看 mcpskill.org 教程 →

配置

{
  "mcpServers": {
    "figma-mcp-go": {
      "command": "npx",
      "args": ["-y", "@vkhanhqui/figma-mcp-go"]
    }
  }
}