#5151st of 82 in Agent frameworks and SDKs
agentscope-java
Java framework for building multi-tenant agents with sandboxes and tool approvals
- Stars
- 6.2k
- License
- Apache-2.0
- Last commit
- Nov 2026
- Last release
- Oct 2026
- Language
- Java
2,366 on 3 Apr 20266,205 on 30 Oct 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
AgentScope Java 2.0 is a JDK 17+ library for building agents on a ReAct core. Its HarnessAgent layer adds workspace files, layered memory, subagents and context compaction. It streams 32 typed events, gates tool calls with allow, approve or deny rules, and runs tools in local, Docker, Kubernetes or E2B sandboxes. Session state can live in Redis, MySQL or PostgreSQL for stateless scaling.
Who it is for: Java teams building production agents with tenancy, approvals and sandboxing
Strengths
- Three-state tool permission engine with human approval built into the event stream
- Session state in Redis, MySQL or PostgreSQL allows cross-replica recovery
- Sandbox options: local subprocess, Docker, Kubernetes and E2B, with snapshot and resume
- Separate model modules for DashScope, OpenAI, Anthropic, Gemini and Ollama
Weaknesses
- Library only: requires JDK 17+ and Maven, no standalone app or UI
- Quickstart uses DashScope and the docs link Aliyun pages, leaning on that ecosystem
- Some features, such as OpenAI Decisions, are only in 2.1.0-SNAPSHOT builds
- No Docker image or compose file provided
What it needs
- no GPU
- Needs JDK 17+, Maven, Redis (optional), MySQL (optional), PostgreSQL (optional), Docker or Kubernetes (optional sandbox)
- Models: DashScope, OpenAI, Anthropic, Gemini, Ollama
Quick answers
Is agentscope-java open source?
Yes. Its code is on GitHub under the Apache-2.0 license.
Does agentscope-java need a GPU?
No, it runs without one.
Also in Agent frameworks and SDKs
See all 82| Rank | Project | Score |
|---|---|---|
| 1 | ppt-masterAgent skill that turns documents into natively editable PowerPoint decks | 95 out of 100 |
| 2 | openai-agents-pythonPython SDK for multi-agent workflows with handoffs, guardrails and tracing | 90 out of 100 |
| 3 | langchainPython framework for building LLM agents and apps from interoperable components | 89 out of 100 |
| #4 | browser-usePython library that lets an LLM agent control a web browser | 89 out of 100 |
| #5 | tradingagentsMulti-agent LLM framework that simulates a trading firm's analysis process | 89 out of 100 |