Welcome to the AgentScope Sample Agents repository! 🎯 This repository provides ready-to-use Python sample agents built on top of:
The examples cover a wide range of use cases — from lightweight command-line agents to full-stack deployable applications with both backend and frontend.
AgentScope is a multi-agent framework designed to provide a simple and efficient way to build LLM-powered agent applications. It offers abstractions for defining agents, integrating tools, managing conversations, and orchestrating multi-agent workflows.
AgentScope Runtime is a comprehensive runtime framework that addresses two key challenges in deploying and operating agents:
- Effective Agent Deployment – Scalable deployment and management of agents across environments.
- Sandboxed Tool Execution – Secure, isolated execution of tools and external actions.
It includes agent deployment and secure sandboxed tool execution, and can be used with AgentScope or other agent frameworks.
- All samples are Python-based.
- Samples are organized by functional use case.
- Some samples use only AgentScope (pure Python agents).
- Others use both AgentScope and AgentScope Runtime to implement full-stack deployable applications with frontend + backend.
- Full-stack runtime versions have folder names ending with:
_fullstack_runtime
📌 Before running any example, check its
README.mdfor installation and execution instructions.
├── alias/ # Agent to solve real-world problems
├── browser_use/
│ ├── agent_browser/ # Pure Python browser agent
│ └── browser_use_fullstack_runtime/ # Full-stack runtime version with frontend/backend
│
├── deep_research/
│ ├── agent_deep_research/ # Pure Python multi-agent research
│ └── qwen_langgraph_search_fullstack_runtime/ # Full-stack runtime-enabled research app
│
├── games/
│ └── game_werewolves/ # Role-based social deduction game
│
├── conversational_agents/
│ ├── chatbot/ # Chatbot application
│ ├── chatbot_fullstack_runtime/ # Runtime-powered chatbot with UI
│ ├── multiagent_conversation/ # Multi-agent dialogue scenario
│ └── multiagent_debate/ # Agents engaging in debates
│
├── evaluation/
│ └── ace_bench/ # Benchmarks and evaluation tools
│
├── data_juicer_agent/ # Data processing multi-agent system
├── sample_template/ # Template for new sample contributions
└── README.md| Category | Example Folder | Uses AgentScope | Use AgentScope Runtime | Description |
|---|---|---|---|---|
| Data Processing | data_juicer_agent/ | ✅ | ❌ | Multi-agent data processing with Data-Juicer |
| Browser Use | browser_use/agent_browser | ✅ | ❌ | Command-line browser automation using AgentScope |
| browser_use/browser_use_fullstack_runtime | ✅ | ✅ | Full-stack browser automation with UI & sandbox | |
| Deep Research | deep_research/agent_deep_research | ✅ | ❌ | Multi-agent research pipeline |
| deep_research/qwen_langgraph_search_fullstack_runtime | ❌ | ✅ | Full-stack deep research app | |
| Games | games/game_werewolves | ✅ | ❌ | Multi-agent roleplay game |
| Conversational Apps | conversational_agents/chatbot_fullstack_runtime | ✅ | ✅ | Chatbot application with frontend/backend |
| conversational_agents/chatbot | ✅ | ❌ | ||
| conversational_agents/multiagent_conversation | ✅ | ❌ | Multi-agent dialogue scenario | |
| conversational_agents/multiagent_debate | ✅ | ❌ | Agents engaging in debates | |
| Evaluation | evaluation/ace_bench | ✅ | ❌ | Benchmarks with ACE Bench |
| General AI Agent | alias/ | ✅ | ✅ | Agent application running in sandbox to solve diverse real-world problems |
A powerful multi-agent data processing system that leverages Data-Juicer's 200+ operators for intelligent data processing:
- Intelligent Query: Find suitable operators from 200+ data processing operators
- Automated Pipeline: Generate Data-Juicer YAML configurations from natural language
- Custom Development: Create domain-specific operators with AI assistance
- Multiple Retrieval Modes: LLM-based and vector-based operator matching
- MCP Integration: Native Model Context Protocol support
Alias-Agent (short for Alias) is designed to serve as an intelligent assistant for tackle diverse and complicated real-world tasks, providing three operational modes for flexible task execution:
- Simple React: Employs vanilla reasoning-acting loops to iteratively solve problems and execute tool calls.
- Planner-Worker: Uses intelligent planning to decompose complex tasks into manageable subtasks, with dedicated worker agents handling each subtask independently.
- Built-in Agents: Leverages specialized agents tailored for specific domains, including Deep Research Agent for comprehensive analysis and Browser-use Agent for web-based interactions.
Beyond being a ready-to-use agent, we envision Alias as a foundational template that can be adapted to different scenarios.
If you:
- Need installation help
- Encounter issues
- Want to understand how a sample works
Please:
- Read the sample-specific
README.md. - File a GitHub Issue.
- Join the community discussions:
| Discord | DingTalk |
|---|---|
 |
 |
We welcome contributions such as:
- Bug reports
- New feature requests
- Documentation improvements
- Code contributions
See the Contributing for details.
This project is licensed under the Apache 2.0 License – see the LICENSE file for details.
- AgentScope Documentation
- AgentScope Runtime Documentation
- AgentScope GitHub Repository
- AgentScope Runtime GitHub Repository
Thanks goes to these wonderful people (emoji key):
 Weirui Kuang 🚧 💻 👀 📖 |
 Osier-Yi 🚧 💻 👀 📖 |
 DavdGao 🚧 |
 qbc 🚧 |
 Lamont Huffman 💻 |
||
Add your contributions
|
||||||
This project follows the all-contributors specification. Contributions of any kind welcome!