Event Streaming (Avro/Kafka)

Type-safe code generation for Apache Kafka and Avro schemas. No more Object types, no more manual casting, no more runtime surprises.

What is it?

Typr generates type-safe JVM code from Avro schemas (.avsc) and protocols (.avpr). You write schemas, run the generator, and get clean, immutable data classes with typed Kafka producers and consumers.

Before (Standard Avro tooling):

public class OrderPlaced extends SpecificRecordBase {
  private Object orderId;        // Actually a UUID, but typed as Object
  private Object customerId;     // Actually a Long
  public Object get(int field) { ... }  // No type safety
}

Object value = order.get(0);
UUID id = (UUID) value;  // Runtime cast, might fail

After (Typr):

public record OrderPlaced(
    UUID orderId,
    Long customerId,
    BigDecimal totalAmount,
    List<String> items
) implements OrderEvents {
    // Immutable, all fields properly typed
}

UUID id = order.orderId();  // No casting needed

Features

Feature	Description
Immutable Records	Java records, Kotlin data classes, Scala case classes
Typed Producers	producer.send(orderId, new OrderPlaced(...))
Typed Consumers	Handler interface with one method per event type
Multi-Event Topics	Sealed interfaces for topics with multiple event types
Complex Unions	["string", "int", "boolean"] → StringOrIntOrBoolean
Wrapper Types	x-typr-wrapper for type-safe IDs
Precise Types	Decimal10_2 with compile-time constraint validation
Typed Headers	Strongly-typed Kafka headers
RPC Support	Request/reply from .avpr protocols
Framework Integration	Spring Boot, Quarkus, Cats (fs2-kafka)

Supported Languages

• Java 21+ - Records with pattern matching
• Kotlin 2.0+ - Data classes with nullable types
• Scala 3 - Case classes with Option types

Wire Formats

Format	Description
Avro + Confluent	Binary Avro with Schema Registry
Avro	Plain binary Avro
JSON	Jackson, Circe, or ZIO JSON

Effect Types

Effect	Return Type	Use Case
Blocking	T	Synchronous code
CompletableFuture	CompletableFuture<T>	Async Java
Mutiny	Uni<T>	Quarkus
CatsIO	IO[T]	Cats Effect
ZIO	Task[T]	ZIO

Next Steps

• Quick Start - Get up and running
• Schema Types - Records, enums, unions
• Kafka Producers - Type-safe event publishing
• Kafka Consumers - Type-safe event handling