A Model Context Protocol (MCP) server for performing Kafka client operations from AI assistants.

mcp-kafka provides a bridge between AI assistants and Apache Kafka, allowing them to interact with Kafka clusters through the Model Context Protocol. This tool enables AI assistants to create, manage, and interact with Kafka topics and messages directly.

Here is a short video demo on youtube

The mcp-kafka server provides the following Kafka operations:

• Create Topic: Create a new Kafka topic with configurable partitions and replication factor
• List Topics: Get a list of all available Kafka topics in the cluster
• Delete Topic: Remove an existing Kafka topic
• Describe Topic: Get detailed information about a specific topic, including partition details
• Produce Message: Send messages to a Kafka topic with support for message keys and headers
• Consume Messages: Read messages from a Kafka topic with configurable timeout

• Go 1.24 or higher
• A running Kafka cluster (default connection: localhost:9092)

1. Clone the repository:

   git clone https://github.com/kanapuli/mcp-kafka.git
   cd mcp-kafka

2. Build the application:

   make build

3. Optionally, build for a specific platform:

   make build GOOS=darwin GOARCH=arm64

To use mcp-kafka with Claude Desktop:

1. Build the application as mentioned above.

2. Place the executable in a location included in your system PATH or in a dedicated tools directory.

3. Please follow this Claude Desktop Tool Installation Guide to install the tool.

The JSON to be added to the claude_desktop_config.json file is as follows, Use the right username and password for SASL login. Leave them empty for PLAINTEXT authentication.

{
  "mcpServers": {
    "kafka": {
	    "command": "/Your-mcp-kafka-executable-path/mcp-kafka-darwin-arm64",
      "args": [
          "--bootstrap-servers=localhost:9092",
          "--consumer-group-id=mcp-kafka-consumer-group",
          "--username=",
          "--password="
        ],
      "env": {}
    }
  }
}

The mcp-kafka tool accepts the following configuration parameters:

These parameters will be automatically derived from your Natural language message to the LLM.

The following flags should be used to configure the Kafka client:

--bootstrap-servers=localhost:9092
--consumer-group-id=mcp-kafka-consumer-group
--username='your_sasl_username'
--password='your_sasl_password'

NOTE: Currently, SASL_PLAINTEXT is supported along with PLAINTEXT authentication. SASL_SSL is not supported.

Contributions are welcome! Please feel free to submit a Pull Request.

MIT License

• Built using the MCP Golang library