Figma Context MCP
Give your coding agent access to your Figma data. Implement designs in any framework in one-shot.
作者提供的说明(英文原文)
Grabbit gives AI agents eyes on the web. Send a public URL, get a pixel-perfect hosted image.
这三项永远不会合并成一个分数:意见、测量和人气是不同的东西。
我们还没有为它写通俗的介绍。下面是我们从注册中心读取到的事实。
选择你使用的 AI 应用,然后复制设置。
你需要
你用的是哪个应用?
在上方选择你的应用。我们会记住它,下次按钮会显示“安装到 …”。
适用于网页版、桌面版和手机版 Claude:
https://mcp.grabbit.live在终端中运行这条命令
claude mcp add --transport http screenshots https://mcp.grabbit.live粘贴到项目中的 .cursor/mcp.json,或 ~/.cursor/mcp.json
{
"mcpServers": {
"screenshots": {
"url": "https://mcp.grabbit.live"
}
}
}浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。
粘贴到项目中的 .vscode/mcp.json
{
"servers": {
"screenshots": {
"type": "http",
"url": "https://mcp.grabbit.live"
}
}
}浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。
ChatGPT 只能连接互联网上的服务器:
https://mcp.grabbit.live菜单名称可能因套餐和版本不同而略有差异。
你的选择通过 Cookie 保存在这台设备上,无需账户。
这些设置是根据作者在注册中心的条目生成的。如果 README 的写法不同,请以 README 为准。
我们还没有读取这个服务器的工具列表;我们会轮流检查各个端点。
我们检查的是任何人都能看到的信息;我们没有阅读代码。不适用的检查项不计入总数。 · 检查于 2026年11月7日
注册名称已通过域名 grabbit.live 验证(注册中心会核实域名所有权)
我们还没有从 GitHub 读取这个仓库
我们还没有从 GitHub 读取这个仓库
尚未检查;我们会轮流检查各个端点
尚未检查;我们会轮流检查各个端点
尚未检查;我们会在检查端点时读取工具列表
与注册中心里这个分类中所有活跃的服务器比较。只有当发布者拥有该仓库时,星标数才计入。
它在 1 项适用的基础检查中通过了 1 项。我们没有阅读代码。在添加密钥或令牌之前,请先看看这个服务器要求哪些权限,并使用只具备所需权限的密钥。
注册中心没有列出必需的密钥,不过有些服务器会在首次使用时要求你登录。
它在作者的云端运行。通过互联网连接,无需安装,但你发送的内容会经过作者的服务器。
GitHub 仓库最后更新于–,最新版本(1.0.2)于39天前进入注册中心。
支持 MCP 的应用,例如 Claude Desktop、Claude Code、Cursor 和 VS Code。它通过 URL 连接,所以接受远程 MCP 服务器的应用也可以使用,例如开启开发者模式的 ChatGPT。
同一分类中有:Figma Context MCP, Claude Real Video, Figma MCP Server
同一分类中其他作者的服务器,按星标数排序。
如何看卡片:★ GitHub 星标i三大支柱:评价 · 检查 · 使用情况i更新 = 最新版本发布到注册中心的时间i
Give your coding agent access to your Figma data. Implement designs in any framework in one-shot.
Let any LLM watch a video locally — and search everything it has ever watched.
The Figma MCP server brings Figma design context directly into your AI workflow.
按原文从作者的仓库复制。PopNx MCP 没有审查或背书其内容。HTML、图片和脚本已被移除。链接会离开本站。
Screenshots as a service, built for AI agents
Point your agent at a URL and get back a hosted image, with no local headless browser to run. Grabbit handles cookie and consent banners, JavaScript-heavy and full-page captures, and many sites that block a local browser.
| Endpoint | https://mcp.grabbit.live (Streamable HTTP) |
| Auth | OAuth 2.1 (claude.ai connectors work out of the box) or a Bearer API key |
| Registry name | live.grabbit/screenshots |
| Setup guides | https://grabbit.live/agents |
| Tool | What it does |
|---|---|
grab | Screenshot any public URL and get a hosted image back (inline image + remaining credits) |
get_grab | Fetch a previous capture by id |
list_grabs | Browse your recent captures |
get_usage | Credits, plan, and 30-day usage |
get_pricing | Pricing plus an honest, dated comparison against ScreenshotOne, Urlbox, Browserless, and others |
Most agents can already drive a headless browser. It works on simple pages and quietly fails on the ones that matter: cookie and consent walls, bot detection, login and paywall gates, and heavy client-rendered pages. Grabbit runs the capture on hosted infrastructure, so the shot comes back clean and the agent needs no browser of its own.
Honest bound: no service captures literally every site. A few hard targets
behind aggressive IP blocking still resist. Private and internal addresses
(including localhost) are blocked by SSRF protection.
claude mcp add --transport http grabbit https://mcp.grabbit.live
Then run /mcp and authenticate in the browser (OAuth, no key to paste). Guide: https://grabbit.live/agents/claude-code
~/.cursor/mcp.json (commit a project-scoped .cursor/mcp.json to share with your team):
{ "mcpServers": { "grabbit": { "url": "https://mcp.grabbit.live" } } }
Guide: https://grabbit.live/agents/cursor
~/.codex/config.toml:
[mcp_servers.grabbit]
url = "https://mcp.grabbit.live" bearer_token_env_var = "GRABBIT_API_KEY"
Guide: https://grabbit.live/agents/codex
Add the hosted MCP server, or wrap the one-line CLI in a skill. Guide: https://grabbit.live/agents/openclaw
Install Grabbit directly from the Cursor Marketplace (when listed), or add manually:
~/.cursor/mcp.json (or project-scoped .cursor/mcp.json):
{ "mcpServers": { "grabbit": { "url": "https://mcp.grabbit.live" } } }
This repo includes Cursor plugin packaging (.cursor-plugin/plugin.json, mcp.json, skill) for marketplace submission.
This repo ships a portable skill folder (skills/grabbit-screenshots/) that teaches agents when and how to use the Grabbit hosted screenshot MCP.
npx skills add BrainGridAI/grabbit-mcp --skill grabbit-screenshots
After public installs, the skill can appear on skills.sh via install telemetry (no separate submit form).
From the repo root:
clawhub skill publish ./skills/grabbit-screenshots --slug grabbit-screenshots --name "Grabbit Screenshots" --version 1.0.0
The skill is published on ClawHub at https://clawhub.ai/braingrid/skills/grabbit-screenshots
openclaw skills install @braingrid/grabbit-screenshots
clawhub install grabbit-screenshots
Hermes can install the same ClawHub skill (ClawHub is a Hermes skill source) and connect MCP at https://mcp.grabbit.live.
npx grabbit.live login # browser sign-in, keys saved locally
npx grabbit.live nyt.com # screenshot a URL, hosted image back npx grabbit.live mcp # print MCP config for your agent
Flat $0.002 per live capture. Prepaid credits that never reset or expire. Test keys return free placeholders so you can wire it up before paying. Full pricing and the comparison table: https://grabbit.live/#pricing
The server.json in this repo is the manifest published to the official MCP Registry. Grabbit's source lives in a private repo; this is the public home for the MCP server's manifest and docs.