Kafka

You run an MCP server that lets a client interact with topics in a Kafka cluster, enabling you to publish and fetch messages through the MCP protocol. This guide shows how to set up and use the Kafka MCP Server for local experimentation and integration with MCP clients.

How to use

You connect your MCP client to the Kafka MCP Server and specify the Kafka bootstrap servers to target. The server runs as a local process and exposes MCP endpoints that translate to Kafka topic operations. To verify your setup, perform a read from a topic and a write to a topic through your MCP client and check the results. If you encounter issues, use the debugging tools in the MCP environment to diagnose configuration or connectivity problems.

How to install

Prerequisites you need before installing:
- Node.js and npm or a compatible runtime environment for MCP runtime execution
- Access to a Kafka cluster or a local Kafka instance for testing
- A local directory where you will store the MCP server code and configuration

Step-by-step setup (manual installation):

1. Clone the Kafka MCP Server repository.
2. Create or locate your Claude desktop configuration file at the following path depending on your OS:

• On MacOS: ~/Library/Application\ Support/Claude/claude_desktop_config.json
• On Windows: %APPDATA%/Claude/claude_desktop_config.json

3. Add the MCP server configuration for Kafka to your Claude config. Use the following structure, replacing placeholders with your actual paths and Kafka bootstrap servers:

{ "mcpServers": { "kafka": { "command": "uv", "args": [ "--directory", "/path/to/repo/mcp-kafka", "run", "mcp-server-kafka", "--kafka-bootstrap-servers", "<comma_separated_bootstrap_servers ex. localhost:9092,otherhost:9092>" ] } } }

Note: You may add additional Kafka connection arguments as needed, such as security protocol, SASL mechanism, username, and password, in the same `args` array as shown above.

4. Install and open the Claude desktop app.
5. Ask Claude to perform a read or write operation to verify the setup, for example by publishing a message to a topic or reading messages from a topic. If something goes wrong, use the MCP debugging tools to investigate.

Additional configuration and notes

Configuration shown above uses a local directory for the MCP server code and passes the Kafka bootstrap servers as an argument. You can extend the command with extra Kafka connection options such as security protocol, SASL mechanism, and credentials as needed for your Kafka cluster.

Security considerations: ensure that any credentials or sensitive connection details are stored securely and not committed to source control. Use environment-specific configuration or secret management where possible.

Troubleshooting tips: check the MCP client logs, verify that Kafka is reachable at the provided bootstrap servers, and ensure the MCP server process has permission to access the specified directory. If the debugging tools indicate a mismatch in topics or permissions, adjust Kafka ACLs accordingly.

Examples and usage notes

• When configuring, keep the bootstrap server list up to date with all brokers in your Kafka cluster to improve reliability and failover handling.

• If you plan to run multiple MCP servers, assign distinct names and separate configuration blocks to avoid conflicts.