agentscope-ai/agentscope-java

中文主页 | Documentation

What is AgentScope Java 2.0?

AgentScope Java 2.0 is a production-ready framework for building distributed, enterprise-grade agents, providing essential abstractions that work with rising model capability and built-in support for long-running, safely-controlled agent execution.

• Event System → A unified event stream with 31 typed events for real-time frontend rendering and human-in-the-loop.
• Permission System → Tool-call gating: allow / require user approval / deny.
• Middleware → AOP-style hook interception for flexibly extending the reasoning-acting loop.
• Workspace & Sandbox → Run tools in isolated environments — local, Docker, Kubernetes, or AgentRun cloud sandbox.
• Multi-Agent Orchestration → Multiple subagent definition patterns with agent_spawn / agent_send and real-time event forwarding.
• Distributed Deployment → True distributed session and memory management (Redis / MySQL / PostgreSQL / OSS / COS) with cross-replica session recovery.
• Agent Evolution → Provides Agent observability and auditing, Agent evaluation and experimentation, and Agent asset management and continuous optimization.

News

• [2026-08] AgentScope Service: An agent Control Plane and Dashboard for agent observation, registration and orchestration, designed to be compatible with AgentScope, LangChain, ADK, and Claude / Qoder, etc.
• [2026-07] v2.0.0 GA: First production-ready release! Dual-layer agent architecture, event stream, permission system, middleware, workspace sandbox, multi-agent orchestration, and distributed deployment all ready. Docs | Release Notes
• [2026-07] v2.0.0-RC5: Model provider modularization; unified DataBlock multimodal support; native structured output; Channel IM integration; Tencent Cloud COS state persistence. Release Notes
• [2026-06] v2.0.0-RC4: Async tool execution and scheduled wakeup dispatching; subagent cross-replica routing and session recovery. Release Notes
• [2026-06] v2.0.0-RC3: Unified call() / streamEvents() execution core; AgentResultEvent, CustomEvent, HintBlockEvent new event types. Release Notes
• [2026-06] v2.0.0-RC2: Fully stateless agent; one-line DistributedBackend; subagent event forwarding; Channel module (DingTalk, Feishu, WeCom); A2A / AG-UI protocol support. Release Notes
• [2026-05] v2.0.0-RC1: First 2.0 RC — harness engineering, enterprise distributed deployment, event stream / message model / middleware / HITL full redesign. Release Notes
• [2026-05] v1.1.0: Introduced agentscope-harness and HarnessAgent; Workspace (persona / memory / skills / subagents); pluggable filesystem (local / shared store / sandbox); session persistence with cross-process resume; layered memory and context compaction; declarative subagent orchestration. Release Notes

Community

Welcome to join our community on

| Discord | DingTalk | WeChat | |--------------------------------------------------------------|-----------------------------------------------------------------------|--------------------------------------------------------------| | | | |

Quickstart

Installation

AgentScope Java requires JDK 17 or higher.

Maven

    io.agentscope
    agentscope-harness
    2.0.1

Model providers are shipped as separate extension modules in 2.0. Add the one you need — for example, DashScope:

    io.agentscope
    agentscope-extensions-model-dashscope
    2.0.1

Other options: agentscope-extensions-model-openai, agentscope-extensions-model-anthropic, agentscope-extensions-model-gemini, agentscope-extensions-model-ollama. See the Model docs for details.

If you only need a bare ReActAgent without workspace / persistence / sandbox, depend on agentscope-core alone.

Hello AgentScope!

Start your first agent with AgentScope Java 2.0:

import io.agentscope.core.agent.RuntimeContext;
import io.agentscope.core.message.UserMessage;
import io.agentscope.harness.agent.HarnessAgent;
import java.nio.file.Paths;
public class FirstAgent {
    public static void main(String[] args) {
        HarnessAgent agent = HarnessAgent.builder()
                .name("assistant")
                .sysPrompt("You are a helpful AI assistant.")
                // ModelRegistry resolves the string and reads the matching
                // API-key env var (e.g. OPENAI_API_KEY or DEEPSEEK_API_KEY)
                // automatically.
                // Examples: "openai:gpt-4.1", "openai:o3",
                // "deepseek:deepseek-v4-flash", "dashscope:qwen-plus",
                // "anthropic:claude-sonnet-4-7", "ollama:llama3"
                .model("dashscope:qwen-plus")
                // Or pass a ChatModel object directly:
                // .model(OpenAIChatModel.builder().model("gpt-4.1").build())
                .workspace(Paths.get(".agentscope/workspace"))
                .build();
RuntimeContext ctx = RuntimeContext.builder()
                .sessionId("demo").userId("alice").build();
// Blocking call
        agent.call(new UserMessage("Hello!"), ctx).block();
// Or stream events for real-time UI rendering
        agent.streamEvents(new UserMessage("Summarize today in three bullets."), ctx)
                .doOnNext(event -> {
                    switch (event.getType()) {
                        case TEXT_BLOCK_DELTA -> System.out.print(
                                ((io.agentscope.core.event.TextBlockDeltaEvent) event).getDelta());
                        case TOOL_CALL_START -> System.out.println(
                                "\n[tool] " + ((io.agentscope.core.event.ToolCallStartEvent) event).getToolCallName());
                        default -> { }
                    }
                })
                .blockLast();
    }
}

AgentScope Service

AgentScope Service — an Agent Control Plane built on AgentScope Harness that provides: + Control Plane. It provides agent registration, discovery, and distributed coordination services for every agent in the enterprise. It works with mainstream agent runtimes including AgentScope, LangChain, ADK, and Claude / Qoder, giving you a single place to inspect agent metrics and operate on live sessions — for example, compressing session context. + Managed Agents Platform Built on the AgentScope Harness runtime, managed agents lets you run multiple Agents on one fully-managed platform under unified operations. The platform hosts Harness capabilities, while tool execution can be delegated to a sandbox that you control. + Agent Teams Agents registered with AgentScope Service can be assembled into one or more Teams, whether the Agent is a self-hosted AgentScope runtime or a low-code Managed Agent Harness runtime, Agents can be orchestrated together to tackle more complex work.

Key Design

AgentScope Java 2.0 is a major step up from a "build an agent" toolkit toward a complete platform for running agents in production. The improvements fall into three focus areas:

1 · Harness Engineering — built for long-running, complex tasks

A bare ReAct loop solves one reasoning turn. HarnessAgent layers engineering infrastructure on top of ReActAgent via Middleware and Toolkit channels — the reasoning core stays untouched, capabilities layer on:

• Self-evolution & skill repository — successful patterns auto-save as Markdown skills, shared across sessions and loaded on demand
• Layered memory — in-context conversation + agent-curated MEMORY.md + on-disk fact log, with auto-compaction to bound the prompt
• Sub-agents — declare child agent specs in Markdown, agent_spawn / agent_send at runtime, sync or background delegation
• Auto context management — structured compaction preserves goals / state / findings / next steps; oversized tool results offload to disk
• Plan Mode — read-only planning state for long tasks, plan files persist and drive execution

2 · Enterprise-grade distributed deployment

Production agents must serve many tenants, run untrusted code safely, and survive rolling restarts. AgentScope 2.0 is built for stateless horizontal scaling:

• Multi-tenant isolation — session / user / agent / org dimension isolation, RuntimeContext keys flow through workspace paths, KV namespaces, and sandbox state slots
• Secure sandbox — local subprocess / Docker / Kubernetes / E2B cloud sandbox, with snapshot and resume
• Permission control — three-state engine (allow / approve / deny), sensitive tools require HITL approval
• Session recovery — AgentStateStore (in-memory / JSON file / MySQL / Redis / PostgreSQL) backs zero-downtime rolling deploys

3 · Foundation framework — a leaner, more developer-friendly core

Messages, events, and the extension model are smaller, more orthogonal — HITL and event streaming are part of how the framework runs, not add-ons:

• Event stream — 31 typed events covering model calls, text deltas, tool execution, and user confirmations in real time
• Message model — text / files / images / audio / video / tool results unified into ContentBlock, role-strict validation at construction
• Middleware — onAgent / onReasoning / onActing / onModelCall / onSystemPrompt five stages replace v1's flat hooks
• HITL first class — confirm tool arguments, approve sensitive actions, hand off to external systems, agent pauses and resumes exactly

For the complete architecture overview, see the documentation.

Contributing

We welcome contributions from the community! Please refer to our CONTRIBUTING.md for guidelines on how to contribute.

License

AgentScope is released under Apache License 2.0.

Contributors

All thanks to our contributors: