Kafka MCP Server

A Message Context Protocol (MCP) server that integrates with Apache Kafka to provide publish and consume functionalities for LLM and Agentic applications.

Overview

This project implements a server that allows AI models to interact with Kafka topics through a standardized interface. It supports:

• Publishing messages to Kafka topics

• Consuming messages from Kafka topics

Related MCP server: KafkaIQ

Prerequisites

• Python 3.8+

• Apache Kafka instance

• Python dependencies (see Installation section)

Installation

1. Clone the repository:

   git clone <repository-url>
   cd <repository-directory>

2. Create a virtual environment and activate it:

   python -m venv venv
   source venv/bin/activate  # On Windows, use: venv\Scripts\activate

3. Install the required dependencies:

   pip install -r requirements.txt

   If no requirements.txt exists, install the following packages:

   pip install aiokafka python-dotenv pydantic-settings mcp-server

Configuration

Create a .env file in the project root with the following variables:

# Kafka Configuration
KAFKA_BOOTSTRAP_SERVERS=localhost:9092
TOPIC_NAME=your-topic-name
IS_TOPIC_READ_FROM_BEGINNING=False
DEFAULT_GROUP_ID_FOR_CONSUMER=kafka-mcp-group

# Optional: Custom Tool Descriptions
# TOOL_PUBLISH_DESCRIPTION="Custom description for the publish tool"
# TOOL_CONSUME_DESCRIPTION="Custom description for the consume tool"

Usage

Running the Server

You can run the server using the provided main.py script:

python main.py --transport stdio

Available transport options:

• stdio: Standard input/output (default)

• sse: Server-Sent Events

Integrating with Claude Desktop

To use this Kafka MCP server with Claude Desktop, add the following configuration to your Claude Desktop configuration file:

{
    "mcpServers": {
        "kafka": {
            "command": "python",
            "args": [
                "<PATH TO PROJECTS>/main.py"
            ]
        }
    }
}

Replace <PATH TO PROJECTS> with the absolute path to your project directory.

Project Structure

• main.py: Entry point for the application

• kafka.py: Kafka connector implementation

• server.py: MCP server implementation with tools for Kafka interaction

• settings.py: Configuration management using Pydantic

Available Tools

kafka-publish

Publishes information to the configured Kafka topic.

kafka-consume

consume information from the configured Kafka topic.

• Note: once a message is read from the topic it can not be read again using the same groupid

Create-Topic

Creates a new Kafka topic with specified parameters.

• Options:

  • --topicName of the topic to create

  • --partitionsNumber of partitions to allocate

  • --replication-factorReplication factor across brokers

  • --config(optional) Topic-level configuration overrides (e.g., retention.ms=604800000)

Delete-Topic

Deletes an existing Kafka topic.

• Options:

  • --topicName of the topic to delete

  • --timeout(optional) Time to wait for deletion to complete

List-Topics

Lists all topics in the cluster (or filtered by pattern).

• Options:

  • --bootstrap-serverBroker address

  • --pattern(optional) Regular expression to filter topic names

  • --exclude-internal(optional) Exclude internal topics (default: true)

Topic-Configuration

Displays or alters configuration for one or more topics.

• Options:

  • --describeShow current configs for a topic

  • --alterModify configs (e.g., --add-config retention.ms=86400000,--delete-config cleanup.policy)

  • --topicName of the topic

Topic-Metadata

Retrieves metadata about a topic or the cluster.

• Options:

  • --topic(If provided) Fetch metadata only for this topic

  • --bootstrap-serverBroker address

  • --include-offline(optional) Include brokers or partitions that are offline