• Deployment Infrastructure: Built-in services for session management, memory, and sandbox environment control
• Sandboxed Tool Execution: Isolated sandboxes ensure safe tool execution without system compromise
• Developer Friendly: Simple deployment with powerful customization options
• 🚧 Framework Agnostic: Not tied to any specific framework. Works seamlessly with popular open-source agent frameworks and custom implementations
• 🚧 Observability: Comprehensive tracing and monitoring for runtime operations

Welcome to join our community on

DingTalk

• 🚀 Quick Start
• 🔌 Agent Framework Integration
• 🏗️ Deployment
• 🤝 Contributing
• 📄 License

• Java 17+

mvn clean install -DskipTests

<dependency>
	<groupId>io.agentscope</groupId>
	<artifactId>agentscope-runtime</artifactId>
	<version>0.1.0-SNAPSHOT</version>
</dependency>

The source code of the following example can be found in the examples directory.

This example demonstrates how to delegate a simple Spring AI Alibaba ReactAgent using AgentScope Runtime Agent.

public static void main(String[] args) {
	try {
		// Create Spring AI Alibaba ReactAgent Builder
		Builder builder = ReactAgent.builder()
			.name("saa_agent")
			.model(chatModel);

		// Create runtime agent proxy using the ReactAgent Builder
		SaaAgent saaAgent = SaaAgent.builder()
			.name("saa_agent_proxy")
			.description("An agent powered by Spring AI Alibaba ReactAgent")
			.reactAgentBuilder(builder)
			.build();

		// Create Runner with the SaaAgent
		Runner runner = new Runner(saaAgent, contextManager);

		// Create AgentRequest
		AgentRequest request = createAgentRequest("Hello, can you tell me a joke?");

		// Execute the agent and handle the response stream
		Flux<Event> eventStream = runner.streamQuery(request);

		eventStream.subscribe(
			event -> handleEvent(event),
			error -> System.err.println("Error occurred: " + error.getMessage()),
			() -> System.out.println("Conversation completed.")
		);

		// Wait a bit for async execution (in real applications, you'd handle this properly)
		Thread.sleep(5000);
	} catch (Exception e) {
		e.printStackTrace();
	}
}

Developers can tell the agent to use a specific tool and the execution of the tool will be delegated to the sandbox managed by AgentScope Runtime.

SaaAgent saaAgent = SaaAgent.builder()
	.name("saa_agent_proxy")
	.tools(List.of("run_python"))
	.description("An agent powered by Spring AI Alibaba ReactAgent.")
	.reactAgentBuilder(builder) //
	.build();

Runner runner = new Runner(saaAgent, contextManager);
AgentRequest request = createAgentRequest("What is the 8th number of Fibonacci?");
Flux<Event> eventStream = runner.streamQuery(request);

AgentScope Runtime Java implementation currently can automatically load Agents developed using Spring AI Alibaba. More agent framework integrations coming soon!

AgentScope Java can expose Agents on a port in the form of standard A2A protocol.

Change the port through the application.yml file:

server:
  port: 8090

Run io.agentscope.runtime.engine.deployer.LocalDeployer to start the A2A server.

We welcome contributions from the community! Here's how you can help:

• Use GitHub Issues to report bugs
• Include detailed reproduction steps
• Provide system information and logs

• Discuss new ideas in GitHub Discussions
• Follow the feature request template
• Consider implementation feasibility

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

For detailed contributing guidelines, please see CONTRIBUTE.