# Use AI Tools with Confluent Cloud You can use AI assistants and coding agents to manage Confluent Cloud resources and build streaming applications. Confluent provides MCP servers that connect AI assistants to your environments, clusters, topics, and connectors, and agent skills that guide you through building production-ready Kafka applications. ## Manage Confluent Cloud with MCP The Model Context Protocol (MCP) is an open standard that connects AI assistants to external tools and data sources. Confluent provides managed MCP servers and an open-source MCP server that let AI assistants interact with your Confluent Cloud resources through natural language. With the MCP servers, your AI assistant can list environments, inspect topics, consume messages, debug connectors, query metrics, write Flink SQL, and more. All actions respect your existing Confluent Cloud permissions. The managed servers provide read-only access, while the open-source server supports both read and write operations. Both servers use your existing credentials and can only access resources those credentials authorize. ### Choose an MCP server Confluent provides managed and open-source MCP servers with different hosting models and capabilities. The following table compares the two options to help you choose. | Feature | Managed MCP servers | Open-source MCP server | |--------------------------------------------|--------------------------------------------|----------------------------------------------| | Hosting | Hosted by Confluent | Self-hosted on your machine | | Transport | Streamable HTTP | stdio, HTTP, SSE | | Confluent Cloud support | Yes | Yes | | Local Confluent Platform and Kafka support | No | Yes | | Kafka topics | Read-only | Read and write | | Flink SQL | Not available | Create, list, read, and delete statements | | Connectors | Read-only inspection and debugging | Create, read, and delete | | Schema Registry | Read-only | List and delete schemas | | Tableflow | Not available | Create, read, update, and delete | | Metrics | Query metrics | Query metrics | | Authentication | Global, Cloud, or Flink API keys | API keys or OAuth | | Setup | Configure endpoint URLs in your MCP client | Install with `npx` and configure a YAML file | For details on setting up each server, see the following pages: - [Access Confluent Cloud with the Managed MCP Servers](managed-mcp-server.md#ai-tools-managed-mcp) - [Build with the Open-Source MCP Server](open-source-mcp-server.md#ai-tools-oss-mcp) ### Supported MCP clients You can use the MCP servers with any MCP-compatible AI assistant, such as the following: - [Claude Code](https://docs.anthropic.com/en/docs/claude-code) - [Claude Desktop](https://claude.ai/download) - [Cursor](https://www.cursor.com/) - VS Code with [GitHub Copilot](https://github.com/features/copilot) - [Windsurf](https://windsurf.com/) - [Goose CLI](https://block.github.io/goose/) - [Gemini CLI](https://github.com/google-gemini/gemini-cli) The managed MCP servers support clients that use streamable HTTP transport. The open-source MCP server supports stdio, HTTP, and SSE transports. ### Relationship to the Real-Time Context Engine The MCP servers described on this page help you manage the the following Confluent resources: environments, clusters, topics, connectors, and metrics. The [Real-Time Context Engine](../real-time-context-engine/overview.md#real-time-context-engine) also uses MCP, but serves a different purpose. It delivers real-time business data from your Kafka topics to AI agents and applications. The two are complementary. You can use the MCP servers to set up and monitor your Confluent Cloud infrastructure, and the Real-Time Context Engine to connect your AI applications to the data flowing through that infrastructure. ## Build streaming applications with agent skills Agent skills are AI-powered workflows that guide coding assistants through building production-ready streaming applications step by step. You get code generation, decision-making support, and guidance on Schema Registry integration, security configuration, and error handling. You can use agent skills to build Kafka producers and consumers, Kafka Streams applications, CDC pipelines, and Schema Registry migration plans. Agent skills work with Claude Code, Cursor, GitHub Copilot, and any AI coding assistant that supports the [Agent Skills Specification](https://agentskills.io/specification). For details, see [Build Streaming Applications with Agent Skills](agent-skills.md#ai-tools-agent-skills).