MCP Servers
The open registry for Model Context Protocol servers. Find the right tools, resources, and prompts for your AI agents — filtered by category, transport, or use case.
Servers
160
Tools
465
Categories
11
Contributors
142
Confluent's open-source MCP server that connects AI assistants to Confluent Cloud, Confluent Platform, and standalone Apache Kafka deployments. Provides tools to manage Kafka topics and connectors, work with Schema Registry, and run Flink SQL statements through natural language, helping teams operate streaming data platforms from an MCP client.
Honeycomb's MCP server that lets AI assistants query and analyze observability data, including events, traces, alerts (triggers), and boards. Agents can run queries against datasets, inspect columns and schemas, and cross-reference production behavior with the codebase to investigate incidents. Connects to Honeycomb via API key or OAuth.
Pulumi's official MCP server for AI-assisted Infrastructure as Code. Wraps the Pulumi Automation and Cloud APIs so agents can preview and deploy stacks, read stack outputs, inspect resources, and look up provider/resource schemas from the Pulumi Registry. Helps developers codify cloud architectures and review infrastructure diffs from within an AI coding assistant.
Skills vs MCP servers
what's the difference?Skillsthe “what to do”
A skillA reusable, structured prompt/workflow with recommended models, an example prompt, and compatible tools. packages know-how — instructions, an example promptA ready-to-use prompt template that demonstrates how to invoke the skill., and recommended models — so an agent performs a task consistently. Skills add knowledge, not new connections.
MCP serversthe “how to connect”
An MCP serverModel Context Protocol server — a standard way to expose tools, resources, and prompts to AI agents and IDEs. gives an agent new capabilities by connecting it to real systems (databases, APIs, files) over a transportHow the client talks to the server: stdio (local process), SSE, or HTTP streaming.. MCP adds connections and actions, not task instructions.
Rule of thumb: reach for a skill when you need the model to do a task well, and an MCP server when you need it to reach a tool or system. They compose — a skill can rely on tools an MCP server provides.
Built an MCP server?
Submit it to the registry — it's open source and community-maintained.