Cloudivo MCP
Model Context Protocol server that acts as an AI gateway for the Cloudivo ecosystem, exposing tools, resources, and prompts while delegating business behavior to downstream services.
README
Cloudivo MCP
Cloudivo MCP is the Model Context Protocol server that will act as the AI gateway for the Cloudivo ecosystem. It exposes MCP tools, resources, and prompts while delegating business behavior to downstream Cloudivo services.
Vision
The long-term goal is a production-ready MCP server that gives AI clients a consistent interface to Cloudivo capabilities without embedding business logic in the gateway itself.
- MCP owns orchestration, registration, authentication, and protocol boundaries.
- Services own business rules.
- Clients own outbound API communication.
Architecture Overview
The application follows a strict layered flow:
Tool -> Service -> Client -> REST API
Bootstrap is intentionally small and modular:
src/index.tsorchestrates startup.src/server/createServer.tscreates the MCP server.src/server/toolRegistry.tsregisters tools.src/server/resourceRegistry.tsregisters resources.src/server/promptRegistry.tsregisters prompts.src/server/startServer.tsattaches transport and starts the process.
See docs/ARCHITECTURE.md for the detailed architecture reference.
Folder Structure
src/
clients/ External API clients
config/ Centralized runtime configuration
prompts/ MCP prompt definitions
resources/ MCP resource definitions
server/ Bootstrap and registry modules
services/ Business-facing orchestration without transport concerns
tools/ MCP tool registration modules
types/ Shared application types
utils/ Small cross-cutting helpers only
tests/ Unit and registry tests
docs/ Project documentation
See docs/PROJECT_STRUCTURE.md for folder responsibility guidance.
Development Workflow
- Install dependencies with
npm install. - Start local development with
npm run dev. - Build the project with
npm run build. - Run the test suite with
npm run test:run. - Validate TypeScript without emitting files with
npm run typecheck.
When adding new MCP functionality, keep src/index.ts unchanged and register new surfaces through the appropriate registry module.
Build Instructions
Prerequisites:
- Node.js 22 LTS
- npm 10+
Commands:
npm install
npm run build
npm start
The build emits compiled output to dist/.
Testing
Vitest is used for unit and registry tests.
npm testruns Vitest in the default mode.npm run test:runruns the suite once.npm run typecheckvalidates types without generating build output.
See docs/TESTING.md for the test strategy and roadmap.
Documentation
Core project references:
docs/ARCHITECTURE.mddocs/adr/README.mddocs/PROJECT_STRUCTURE.mddocs/DEVELOPMENT_GUIDE.mddocs/CODING_STANDARDS.mddocs/TESTING.mddocs/ENGINEERING_PRINCIPLES.mddocs/DEFINITION_OF_DONE.mddocs/SECURITY.mddocs/OPERATIONS.mddocs/adr/
Roadmap
Near-term priorities:
- Harden configuration, logging, and error handling.
- Add service and client foundations for external Cloudivo systems.
- Expand test coverage across registries, services, and future clients.
See docs/ROADMAP.md for the phased plan.
Future Integrations
- CTE
- MailAPI
- Leinaflow
- Future Cloudivo services
Each integration will follow the same layering rules and will be introduced through dedicated client and service modules.
The first reusable integration boundary is the CTE client in src/clients/cte/CteClient.ts.
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。