@floopfloop/mcp

@floopfloop/mcp

MCP server for the FloopFloop API that lets LLM hosts build, poll, and refine projects from natural language prompts.

Category
访问服务器

README

@floopfloop/mcp

npm version npm downloads CI Node.js Version License: MIT

Model Context Protocol server for the FloopFloop API — lets Claude Desktop, Cursor, Zed, and any other MCP-aware LLM host build, poll, and refine FloopFloop projects on the user's behalf.

Wraps the official @floopfloop/sdk and exposes a curated subset of its surface as MCP tools.

Tools

Tool What it does
list_projects List every project you have access to
get_project Fetch a project by id or subdomain
project_status Cheap status snapshot — safe to poll
create_project Kick off a new build from a natural-language prompt
refine_project Send a refinement message to an existing project
wait_for_live Block until the project reaches a terminal state
check_subdomain Is a given slug available?
suggest_subdomain Generate a friendly slug from a prompt
list_secrets List secret keys on a project (names only)
set_secret Create/overwrite a project secret
remove_secret Delete a project secret
list_library_projects Browse the public project library
clone_library_project Duplicate a library project into the user's account
usage_summary Plan limits + current-period credit / build / storage usage
list_api_keys List the user's API keys (never returns the raw secret)
create_api_key Mint a new API key — raw secret returned ONCE
remove_api_key Revoke an API key by id or name
upload_from_path Read a local file, presign + upload to S3, return an attachment ref
whoami Show the authenticated user

Configuration

Grab an API key: floop keys create mcp-host (via the floop CLI) or the dashboard → Account → API Keys. Business plan required to mint new keys.

The server reads FLOOP_API_KEY from the environment. Optionally override FLOOP_API_URL to point at staging.

Claude Desktop

~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %AppData%\Claude\claude_desktop_config.json (Windows):

{
  "mcpServers": {
    "floopfloop": {
      "command": "npx",
      "args": ["-y", "@floopfloop/mcp"],
      "env": { "FLOOP_API_KEY": "flp_..." }
    }
  }
}

Restart Claude Desktop; the tools show up under the 🔌 icon.

Cursor

~/.cursor/mcp.json (or Cursor Settings → Tools & Integrations → MCP):

{
  "mcpServers": {
    "floopfloop": {
      "command": "npx",
      "args": ["-y", "@floopfloop/mcp"],
      "env": { "FLOOP_API_KEY": "flp_..." }
    }
  }
}

Generic MCP host

Any host that speaks the MCP stdio transport can run the server:

FLOOP_API_KEY=flp_... npx -y @floopfloop/mcp

Install

The package is pulled in automatically via npx -y @floopfloop/mcp in the config snippets above. If you prefer a pinned install:

npm install -g @floopfloop/mcp
which floop-mcp

Usage notes

  • create_project starts a build right away. Follow up with wait_for_live (blocks) or project_status (poll) to know when it's up.
  • wait_for_live defaults to a 10-minute ceiling (bounded by timeoutMs, capped at 30 min). Most builds finish in under two minutes.
  • set_secret / remove_secret are marked destructiveHint: true — hosts may ask the user to confirm before they run.
  • list_secrets only returns names, never values. Secret values cannot be retrieved once written; rotate them by re-setting.
  • On failure, tools return an MCP isError content result rather than tearing down the session, so the host displays the error to the user.

Development

git clone https://github.com/FloopFloopAI/floop-mcp.git
cd floop-mcp
npm install
npm run typecheck
npm run build

FLOOP_API_KEY=flp_... node dist/index.js

To smoke-test the stdio handshake without an LLM host:

printf '%s\n' \
  '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2024-11-05","capabilities":{},"clientInfo":{"name":"smoke","version":"0"}}}' \
  '{"jsonrpc":"2.0","method":"notifications/initialized"}' \
  '{"jsonrpc":"2.0","id":2,"method":"tools/list"}' \
  | FLOOP_API_KEY=flp_dummy node dist/index.js 2>/dev/null

Releasing

.github/workflows/release.yml publishes to npm with provenance whenever a tag matching mcp-v* is pushed. One-time setup before the first release:

  1. Create an automation npm token on the @floopfloop scope (publishes bypass 2FA by design when tokens are marked automation).
  2. Add it as the repo secret NPM_TOKEN:
    gh secret set NPM_TOKEN --repo FloopFloopAI/floop-mcp
    # paste token when prompted — do NOT use --body, which can leak it to
    # shell history.
    
  3. Tag + push:
    npm version 0.1.0-alpha.1 --no-git-tag-version  # if needed
    git tag mcp-v$(node -p "require('./package.json').version")
    git push --follow-tags
    

The workflow typechecks, builds, runs the stdio smoke test, verifies the tag matches package.json, then publishes and cuts a GitHub Release (prerelease, auto-generated notes).

License

MIT

推荐服务器

Baidu Map

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。

官方
精选
JavaScript
Playwright MCP Server

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。

官方
精选
TypeScript
Magic Component Platform (MCP)

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。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

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。

官方
精选
本地
TypeScript
VeyraX

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。

官方
精选
本地
graphlit-mcp-server

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。

官方
精选
TypeScript
Kagi MCP Server

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。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

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。

官方
精选