原型 · 预览版:收录官方注册中心里的全部服务器,每 30 分钟同步一次。评价和账户功能将在后续阶段推出。
PopNx MCP为使用 AI 的人准备的 MCP 服务器 · 用大白话讲清楚

Spark

作者:PlusDocs版本 1.0.2i★ –i在云端运行
注册名称io.github.PlusDocs/spark

安装

概览

作者提供的说明(英文原文)

Build and run a business from your AI assistant, with hosting, payments, email and more built in.

PopNx MCP 的三大支柱

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。

它能帮你做什么

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。

  • 连接方式: 在云端运行
  • 传输方式: Streamable HTTP
  • 分类: 聊天与邮件

安装

选择你使用的 AI 应用,然后复制设置。

安装之前

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:

  1. 打开 Settings → Connectors
  2. 点击 Add custom connector
  3. 粘贴下面的地址,然后点击 Add
  4. 如果要求登录,请登录作者的服务
SettingsConnectors1+ Add2https://…3
https://mcp.spark.site

在终端中运行这条命令

claude mcp add --transport http spark https://mcp.spark.site

在 Cursor 中打开

或者自己复制设置

粘贴到项目中的 .cursor/mcp.json,或 ~/.cursor/mcp.json

{
  "mcpServers": {
    "spark": {
      "url": "https://mcp.spark.site"
    }
  }
}

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。

在 VS Code 中打开

或者自己复制设置

粘贴到项目中的 .vscode/mcp.json

{
  "servers": {
    "spark": {
      "type": "http",
      "url": "https://mcp.spark.site"
    }
  }
}

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:

  1. 打开网页版 ChatGPT → Settings → Connectors (Apps)
  2. 点击 Create 或 Add(部分套餐需要先在 Advanced 中开启 Developer mode)
  3. 粘贴下面的地址,起个名字,然后保存
  4. 如果要求登录,请登录作者的服务
SettingsConnectors1+ Create2https://…3
https://mcp.spark.site

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。

它能做什么

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。

基础检查i

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6日

  • ✓ 通过

    作者与仓库一致

    注册名称已通过 GitHub 账户 PlusDocs 验证,该账户拥有此仓库

  • … 尚未检查

    仍在维护

    我们还没有从 GitHub 读取这个仓库

  • … 尚未检查

    有许可证

    我们还没有从 GitHub 读取这个仓库

  • … 尚未检查

    端点有响应

    尚未检查;我们会轮流检查各个端点

  • … 尚未检查

    使用 OAuth 登录

    尚未检查;我们会轮流检查各个端点

  • … 尚未检查

    工具描述中没有隐藏指令

    尚未检查;我们会在检查端点时读取工具列表

与“聊天与邮件”中的其他服务器比较i

  • 星标数多于此分类中其他 1760 个服务器的 0%
  • 发布到注册中心的时间比此分类中其他 1760 个服务器的 78% 更新
  • 基础检查结果等于或优于此分类中其他 1760 个服务器的 100%

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。

常见问题

Spark 安全吗?

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。

它需要哪些密钥或代码?

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。

它在我的电脑上运行还是在云端?

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。

它还在维护吗?

GitHub 仓库最后更新于–,最新版本(1.0.2)于7天前进入注册中心。

哪些应用可以使用它?

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。

有什么可以替代它?

同一分类中有:Vexa, emailmd, MCP WhatsApp Cloud

替代选择

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

Vexa

作者:Vexa-ai

Meeting bot and transcripts for Google Meet, Teams and Zoom. Live or after, speakers labelled.

在云端运行★ 2,949Apache-2.02个月前更新
安装

emailmd

作者:anypost

Render markdown into email-safe HTML, lint drafts for deliverability problems, and preview emails.

电脑或云端★ 1,385MIT30天前更新
安装

MCP WhatsApp Cloud

作者:codespar

MCP server for WhatsApp Cloud API (Meta) — messages, media, and templates via Graph API

在你的电脑上运行★ 277MIT5个月前更新
安装

作者的 README

打开 README

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。

Spark MCP

Build and run a business from your AI assistant with Spark: sites and apps with hosting, customer accounts, payments, email and marketing built into the platform. This is the official plugin for Spark's remote MCP server, for ChatGPT, Claude, Claude Code, Cursor, Codex and any other MCP client.

You:    Plan a new Spark project with me: a booking site for my pottery classes that takes payments.

Agent: Spark has a plan for Clay Studio Bookings. Shall I create it? You: Yes. Agent: Spark is building it. Your preview: https://app.spark.site/projects/…

What Spark is

Spark is a platform for building and running a business. You describe what you want, and Spark designs it, writes the code and publishes it. A project is the whole business, and what it needs to run is built into the platform:

  • Hosting on a spark.site address or your own domain.
  • Customer accounts: sign-in for your customers, and what each one has bought.
  • Payments through Stripe: one-off sales and subscriptions, paid out to you.
  • Email to your customers.
  • Marketing: a marketing plan for the business and the content to go with it.
  • Internal tools for running it, such as customer and payments dashboards, which Spark builds and changes like the site.

Through this server, your assistant works in a project's chat the way you would in Spark. It sends your requests, follows Spark while it works, brings Spark's questions and approvals back to you, looks at the preview and publishes.

What to expect.

  • A change usually takes a few minutes, and a new project's first build about ten. The assistant waits for Spark and reports what it is doing meanwhile.
  • Spark asks before some actions, such as deleting data. Your assistant shows you the question, and nothing happens until you answer.
  • Messages your assistant sends appear in the project's chat in Spark, marked as sent from that assistant, so you can carry on in either place.
  • Building uses the Build credits of your Spark plan, as it does in Spark.

Requirements

A Spark account, on any plan. The Free plan includes one-time credits for building, and you can publish on a spark.site address. Taking payments and using your own domain need a paid plan. See pricing.

You don't need an API key: you sign in with your Spark account the first time you use the server.

Install

ChatGPT. Open Plugins, select +, then Add custom MCP server. Name it Spark, enter https://mcp.spark.site as the server URL with OAuth authentication, and create it. In a chat, type @ and pick Spark. See OpenAI's guide.

Claude (claude.ai, desktop and mobile). Go to Customize → Connectors, select + Add, then Add custom connector. Name it Spark, enter https://mcp.spark.site, and select Continue, then Add. In a chat, turn it on under + → Connectors. On Team and Enterprise plans, an Owner adds it under Organization settings → Connectors and each member selects Connect. See Spark's guide.

Claude Code. Install the plugin, which adds the MCP server and the build-with-spark skill:

/plugin marketplace add PlusDocs/spark-site-mcp

/plugin install spark@spark

Then ask for a change, or run /spark:build-with-spark add a pricing page to my Harbor Tours site. To add only the MCP server: claude mcp add --transport http spark https://mcp.spark.site

Codex. Add this repository as a plugin marketplace, then install Spark.

Cursor and any other MCP client. Add the server to the client's MCP config:

{

"mcpServers": { "spark": { "type": "streamable-http", "url": "https://mcp.spark.site" } } }

On first use, the client opens Spark's sign-in. Sign in to Spark if you aren't already, then choose Connect.

Tools

ToolWhat it does
whoamiSays which Spark account the server is signed in as, and its organizations.
list_projectsLists your projects with their addresses and whether they have unpublished changes.
get_projectOne project: its site's pages, its internal tools, what Spark is doing and any question it is waiting on, and the setup plan's progress.
plan_projectPlans a new project with you, one question at a time, ending with a plan to agree to. Creates nothing.
create_projectCreates the project you agreed to and starts its first build.
send_messageSends a request to Spark on a project, optionally about one page or internal tool, with up to five files.
get_turnFollows a request: what Spark is doing, its reply when done, and the questions it is waiting on. Waits up to 50 seconds per call.
answer_questionAnswers Spark's question or approval, with the choice you picked.
stop_buildStops Spark's work on a request. The changes made so far stay in the preview.
view_pageShows one screen of a page of the preview, at desktop or phone width, so the assistant can check the work.
attach_filesShows Spark's upload card, where you pick files from your computer to send with your next request.
get_historyWhat happened on a project lately: its publishes and recent requests.
get_creditsYour plan and the credits you have left.
publish_projectPublishes the newest version of a project's site once Spark has finished.

In clients that support MCP Apps, such as Claude and ChatGPT, a project shows as a card. The card follows Spark while it works, shows its questions with their choices, and has Publish, Stop and Open in Spark. In Claude, after a change, it shows a picture of the part of the site that changed. Other clients get the same results as text.

Example prompts

  • "Plan a new Spark project with me: a booking site for my pottery classes that takes payments."
  • "What is Spark working on in my Harbor Tours project?"
  • "Ask Spark to add a pricing page with our three class packages, then check it on a phone."
  • "Ask Spark to email my customers about the new spring classes."
  • "Send Spark this menu and ask it to put it on the home page: https://example.com/menu.pdf"
  • "Publish Harbor Tours."

Notes

  • Tools run as the Spark account that signed in, in that account's organizations.
  • Clients may ask you to confirm each request that changes a project, and each publish.
  • To disconnect an assistant, open Account settings in Spark and go to Connected assistants.
  • Spark hosts the server. This plugin runs no local code.

Links

What's in this repository

PathPurpose
.mcp.jsonMCP server config used by Claude Code, Cursor and Codex
plugin.json, mcp.jsonPortable Agent Plugins manifest and MCP config
.claude-plugin/Claude Code plugin manifest and marketplace entry
.cursor-plugin/Cursor plugin manifest
.codex-plugin/, .agents/plugins/Codex plugin manifest and marketplace entry
skills/build-with-spark/Skill with guidance on using the tools well
assets/Logo

The server's listing in the official MCP Registry is io.github.PlusDocs/spark, published with the server itself.

License

MIT

问 PopNx