> ## Documentation Index
> Fetch the complete documentation index at: https://docs.mage.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Kafka

> How to configure Apache Kafka as a destination in Mage to publish JSON records to a Kafka topic using optional message keys.

![Kafka logo](https://upload.wikimedia.org/wikipedia/commons/thumb/5/53/Apache_kafka_wordtype.svg/2560px-Apache_kafka_wordtype.svg.png)

***

## Overview

Use **Apache Kafka** as a destination in Mage to stream records from your pipeline into Kafka topics in real time. Each record is serialized as a JSON object and sent to the configured `bootstrap_server`.

If a `Key Property` is defined in the Source Schema, Mage will use the corresponding field value as the Kafka message key—enabling partitioned message routing and supporting downstream systems that rely on message keys.

This is ideal for:

* Real-time data streaming
* Event-driven microservices
* Kafka-based data pipelines and messaging queues

***

## Configuration Parameters

| Key                | Description                                                                                         | Default | Required |
| ------------------ | --------------------------------------------------------------------------------------------------- | ------- | -------- |
| `bootstrap_server` | Comma-separated list of Kafka brokers with ports (e.g., `kafka:9093`, `broker1:9092,broker2:9092`). | None    | ✅        |
| `topic`            | Kafka topic to publish messages to.                                                                 | None    | ✅        |
| `api_version`      | *(Optional)* Kafka API version to use. If unspecified, Mage will auto-detect the server version.    | None    | ❌        |

***

## Notes

* Records are serialized as **JSON** using the [Kafka Python client](https://kafka-python.readthedocs.io/en/master/).
* To set Kafka message keys, define a `Key Property` in your source schema.
* You can explicitly set the `api_version` to match your Kafka broker version if needed.
* Compatible with self-hosted Kafka, Confluent Cloud, Amazon MSK, and other Kafka-compatible platforms.

***

## Example Configuration

```yaml theme={"system"}
bootstrap_server: broker1.kafka.internal:9092,broker2.kafka.internal:9092
topic: mage_data_stream
api_version: (0, 10, 1)
```

***

## Related Resources

* [Apache Kafka Documentation](https://kafka.apache.org/documentation/)
* [Kafka Python Client (kafka-python)](https://kafka-python.readthedocs.io/en/master/)
* [Confluent Kafka Topic Design](https://developer.confluent.io/learn/kafka/topic-design/)