MCP 服务器
开放的 Model Context Protocol 服务器注册表。按类别、传输方式或使用场景,为 AI 智能体寻找合适的工具、资源和提示词。
服务器
177
工具
471
类别
11
贡献者
155
Official Heroku MCP server that lets AI agents manage Heroku Platform resources. Supports listing and inspecting apps, scaling dynos, viewing logs, managing config vars and add-ons, and running one-off commands, so deployment and operations tasks can be handled conversationally.
MCP server for the Railway deployment platform. Lets AI agents create projects and services, deploy from repositories, manage environment variables, view deployment logs, and inspect service status. Useful for provisioning backends, databases, and cron jobs and for debugging deploys directly from an AI assistant.
Render's official MCP server for interacting with your Render cloud resources via LLMs. Lets agents list and inspect services, create and manage deploys, read logs and metrics, and query managed Postgres and Key Value instances. Useful for deploying, debugging, and monitoring applications hosted on Render directly from an MCP-compatible assistant.
Skills 与 MCP 服务器
有什么区别?Skills定义“做什么”
Skill 将说明、示例提示词和推荐模型组织在一起,让智能体稳定地完成任务。Skills 增加的是知识,而不是新的连接。
MCP 服务器定义“如何连接”
MCP 服务器通过连接数据库、API 和文件等真实系统,为智能体增加新能力。MCP 提供连接和操作,而不是任务说明。
简单来说:当模型需要把某项任务做好时选择 Skill;当模型需要访问工具或系统时选择 MCP 服务器。两者可以组合使用,Skill 可以依赖 MCP 服务器提供的工具。
开发了 MCP 服务器?
将它提交到由社区维护的开源注册表。