JVM-native runtime

jamjet-runtime-java is the JVM-native durable agent runtime. It ships eight modules on Maven Central and runs durable agents directly on the JVM — no Rust runtime, no REST sidecar, no gRPC bridge.

Use this when:

• Your application is Spring Boot, Quarkus, plain Java, or Kotlin.
• You want durable execution without a sidecar process.
• You're targeting Java 21+ and want virtual threads, not green threads.

Use the Java SDK instead if you want to author workflows in Java that compile to the shared IR and run on the Rust runtime.

Install

Requires Java 21+.

Spring Boot (recommended)

<dependency>
  <groupId>dev.jamjet</groupId>
  <artifactId>jamjet-spring-boot-starter</artifactId>
  <version>0.3.1</version>
</dependency>

Auto-registers DurableAgentBeanPostProcessor. Minimal application.yml:

jamjet:
  storage: in-memory

Plain Java (no Spring)

<dependency>
  <groupId>dev.jamjet</groupId>
  <artifactId>jamjet-runtime-core</artifactId>
  <version>0.3.1</version>
</dependency>
<dependency>
  <groupId>dev.jamjet</groupId>
  <artifactId>jamjet-runtime-instrument</artifactId>
  <version>0.3.1</version>
</dependency>

Authoring a durable agent

Annotate the class with @DurableAgent, the methods with @Checkpoint, and wrap side-effecting work in DurabilityContext.replayOrExecute:

import dev.jamjet.runtime.instrument.DurabilityContext;
import dev.jamjet.runtime.instrument.annotations.Checkpoint;
import dev.jamjet.runtime.instrument.annotations.DurableAgent;

@DurableAgent("research-pipeline")
public class ResearchAgent {

  @Checkpoint("search")
  public String search(String topic) {
    return DurabilityContext.current().replayOrExecute("search", () -> {
      return searchClient.query(topic);
    });
  }

  @Checkpoint("analyze")
  public String analyze(String sources) {
    return DurabilityContext.current().replayOrExecute("analyze", () -> {
      return llmClient.chat(systemPrompt, sources);
    });
  }
}

When a worker crashes between checkpoints, the next worker picks up the run and replayOrExecute returns the previously-recorded result for every completed checkpoint, then executes only the remaining ones.

Spring Boot integration

Adding @Service (or any Spring stereotype) on top of @DurableAgent registers the agent as a normal Spring bean. The DurableAgentBeanPostProcessor discovers it at startup:

@DurableAgent("research-pipeline")
@Service
public class ResearchAgent {
  // …
}

Inject it the way you'd inject any Spring service.

Modules

MCP protocols

jamjet-runtime-protocols ships native MCP server and client adapters. Tools defined on the runtime are exposed as MCP tools without writing protocol glue; remote MCP servers can be added as callable tools the same way.

AgentBoundary receipts

Emit Action Receipts by pairing jamjet-cloud-sdk with the runtime — receipts conform to the open AgentBoundary v0.1 spec and validate against the canonical JSON Schema.

Examples

Four runnable examples ship in the repo:

• crash-recovery — Kill a worker mid-run; another picks up at the next checkpoint.
• real-agent — OpenAI-backed agent with checkpoint persistence.
• spring-boot-comparison — Side-by-side old way (manual durability) vs new way (@DurableAgent).
• latency benchmark — Measures the overhead of the durability layer.

Browse them at jamjet-runtime-java/examples.

Storage backends

The current 0.3.x line supports an in-memory checkpoint store, suitable for single-process deployments and development. The wire protocol for swappable backends is in place; durable backends are tracked on the jamjet-runtime-java issue tracker.

Reference

• GitHub: jamjet-labs/jamjet-runtime-java
• Maven Central: dev.jamjet namespace
• Public landing: jamjet.dev/runtime