gatepass-server
Minimal in-memory Node.js server for the Gatepass gate-control prototype, exposing a REST API and an MCP tool (open_gate) to manage shared gate state.
README
Gatepass Server
For the story behind this project, see the Gatepass README.
Minimal in-memory Node.js server for the Gatepass
Flutter gate-control prototype. It holds a single shared gate state
(idle or open_requested) in memory — no database, no persistence across
restarts — and exposes it over both a small REST API and an MCP tool.
[Remote Control app] --POST /gate/open--> [this server] <--poll GET /gate/status-- [Gate Device app]
^
POST /mcp (open_gate tool)
Endpoints
POST /gate/openGET /gate/statusPOST /gate/ackGET /healthPOST /mcp
Setup
npm install
npm start
Development
npm run dev
Environment
Copy .env.example to .env and set values if needed.
PORTdefaults to3000GATE_API_KEYis optional
If GATE_API_KEY is set, POST /gate/open, GET /gate/status, and POST /gate/ack accept either:
Authorization: Bearer <GATE_API_KEY>x-api-key: <GATE_API_KEY>
If GATE_API_KEY is not set, requests are allowed normally.
Note: POST /mcp does not currently check GATE_API_KEY — the
open_gate MCP tool is reachable without a credential regardless of that
setting. This is a known limitation of the prototype, not an oversight; if
you need the MCP endpoint gated too, add the same requireGateKey
middleware used on the REST routes before deploying anywhere untrusted.
Testing with MCP Inspector
Install and run the Inspector against the local MCP endpoint:
npx @modelcontextprotocol/inspector http://localhost:3000/mcp
If you already have the Inspector installed locally, use its launch command instead and point it at:
http://localhost:3000/mcp
In the Inspector:
- Connect to
http://localhost:3000/mcp - Open the tools list
- Confirm
open_gateis discovered - Invoke
open_gate - Check
GET /gate/statusand confirm it returns{"status":"open_requested"}
Status
This is a working prototype: no database, no request logging beyond
console.log, and no rate limiting. CORS is wide open (all origins
allowed). Treat it as a starting point rather than a production-ready
server.
Production deployments
This repository is deliberately a prototype — it exists to share and open-source the concept.
We also have a hardened version of this architecture designed for production and higher-security deployments, where existing gate or access-control infrastructure can use GSM as the integration layer.
If you have an existing gate deployment that you'd like to make remotely or programmatically accessible, reach out: harvey@blnga.co.zw. We can talk through the security model, live deployment, monitoring, and reporting.
License
MIT — see LICENSE.
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。