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
175
工具
465
类别
11
贡献者
153
AgentMail MCP gives AI agents secure email operations through hosted MCP access and local stdio bridges for agent-native inbox workflows.
Apache Doris' MCP server provides agents with governed access to Doris analytics databases for schema discovery, query assistance, and operational investigation.
Apache IoTDB's MCP server connects agents to time-series data and metadata for operational analysis, industrial telemetry exploration, and query assistance.
Apache SkyWalking's MCP server gives agents access to observability data for tracing, service topology, metrics, logs, and production incident investigation.
Apache Solr's MCP server connects agents to search collections and query workflows for schema discovery, relevance investigation, and search application development.
Buildkite's official MCP server for querying and operating CI/CD pipelines, builds, jobs, artifacts, and organization-level delivery workflows.
Chrome's official MCP server for inspecting and controlling a live browser through Chrome DevTools. It helps agents diagnose UI, network, performance, and runtime issues.
Google Cloud's MCP server for deploying and managing applications on Cloud Run, giving coding agents a guided path from source code to a serverless production service.
Dagster's MCP integration enables agents to inspect and operate data assets, jobs, schedules, and runs in modern data platforms.
The official MCP reference and test server demonstrates tools, resources, prompts, sampling, and protocol capabilities for MCP client development.
Firebase's MCP integration exposes Firebase project context and development workflows to coding agents for building, testing, and operating Firebase applications.
Fivetran's MCP server helps agents inspect data connectors and sync health, diagnose broken connections, and answer operational questions about managed ELT pipelines.
Google's official MCP server for the Google Ads API, exposing resources and tools that let agents analyze advertising accounts and support campaign operations.
GrowthBook's official MCP server for working with feature flags, experiments, and metrics so agents can safely support product experimentation workflows.
Harness' official MCP server for connecting agents to software delivery workflows, including deployment pipelines, services, environments, and delivery insights.
InfluxData's MCP server connects agents to InfluxDB 3 for SQL queries, schema exploration, and time-series database operations.
JetBrains' official MCP server for working with IntelliJ-based IDEs and Android Studio, enabling agents to inspect projects and use IDE-aware development capabilities.
JFrog's MCP server lets agents work with Platform services for artifact repositories, build information, release lifecycle management, and software supply-chain workflows.
Microsoft's MarkItDown MCP integration converts documents from HTTP, file, and data URIs into Markdown for agent-ready analysis and retrieval.
Google's MCP Toolbox provides secure, ready-made and custom tools for database queries, schema operations, and semantic search in agent workflows.
Medusa's MCP integration connects agents to commerce data and administrative operations for storefront, catalog, order, and customer workflows.
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。