AgentScope

Highlight

Agent

• Support structured output (#2150)
• Support environment information injection (#2134)

Workspace

• Support OpenSandbox (#1953), Daytona (#1943), K8s (#1933), Bubblewrap (#2051)

Tool

• Support PowerShell for Windows (#2132)

RAG

• Support MongoDB (#2008), Elasticsearch (#2129)
• Support Excel (#2026) and Word (#2025) parser

Agent as Service

• Support SQLAlchemy based storage (#2029)

• Support sharing resources across users as a group or org (#1998)

Model

• Add new Dashscope models (#2073), Kimi K3 (#2141)
• Implement Gemini TTS API (#1879)

What's Changed

• fix(gemini): fix gemini schema containing null by @qbc2016 in https://github.com/agentscope-ai/agentscope/pull/2020
• feat(workspace): add K8sWorkspace with Pod/PVC lifecycle, tar-stream backend and MCP gateway support by @XiuShenAl in https://github.com/agentscope-ai/agentscope/pull/1933
• feat(rag): add MongoDB vector store backend Closes by @xu-suping in https://github.com/agentscope-ai/agentscope/pull/2008
• fix(formatter): include tool_name in ollama tool result messages by @nolanchic in https://github.com/agentscope-ai/agentscope/pull/2006
• feat(workspace): add OpenSandbox backend by @lumman0 in https://github.com/agentscope-ai/agentscope/pull/1953
• fix(formatter): drop empty text blocks for anthropic by @nolanchic in https://github.com/agentscope-ai/agentscope/pull/2007
• feat(rag): add excel parser by @qbc2016 in https://github.com/agentscope-ai/agentscope/pull/2026
• feat(rag): add word parser by @qbc2016 in https://github.com/agentscope-ai/agentscope/pull/2025
• feat(service): support to share resource (credential, agent, knowledge bases) across different users as a group or org by @DavdGao in https://github.com/agentscope-ai/agentscope/pull/1998
• feat(workspace): add daytona workspace backend by @BreezyB1n in https://github.com/agentscope-ai/agentscope/pull/1943
• fix(workspace): restore default glob helper path by @dailingtao in https://github.com/agentscope-ai/agentscope/pull/2056
• fix: change max_tokens to max_completion_tokens in openai chat model by @qbc2016 in https://github.com/agentscope-ai/agentscope/pull/2065
• fix(workspace): fix OpenSandbox bootstrap and state filtering by @jwx0925 in https://github.com/agentscope-ai/agentscope/pull/2046
• refactor(webui): update tools rendering in the web ui by @DavdGao in https://github.com/agentscope-ai/agentscope/pull/2072
• feat(dashscope): add new models for dashscope by @qbc2016 in https://github.com/agentscope-ai/agentscope/pull/2073
• fix(webui): fix rendering errors in Read/Write/Edit tools caused by incomplete JSON during tool call streaming or interruption by @DavdGao in https://github.com/agentscope-ai/agentscope/pull/2075
• fix(rag): adapt MilvusLiteStore to milvus-lite 3.1.0 COSINE distance semantics by @qbc2016 in https://github.com/agentscope-ai/agentscope/pull/2089
• chore(deps): restrict MCP dependency version to below 2.0.0 by @qbc2016 in https://github.com/agentscope-ai/agentscope/pull/2091
• fix(model): preserve OpenAI Responses reasoning history replay by @qbc2016 in https://github.com/agentscope-ai/agentscope/pull/2071
• fix(app): use cursor-based pagination for list_messages by @DavdGao in https://github.com/agentscope-ai/agentscope/pull/2081
• fix(webui): scroll newly loaded sessions to bottom by @JanCong in https://github.com/agentscope-ai/agentscope/pull/2100
• fix(mem0): use separate LLM instances in OSS example by @Osier-Yi in https://github.com/agentscope-ai/agentscope/pull/2078
• chore(version): bumping version to 2.0.4.post1 by @qbc2016 in https://github.com/agentscope-ai/agentscope/pull/2104
• fix(agent): preserve multi-tool pairs during context compression by @qiuC123 in https://github.com/agentscope-ai/agentscope/pull/2093
• fix(tool): reject mutating find commands as read-only by @VectorPeak in https://github.com/agentscope-ai/agentscope/pull/2004
• perf(prompt)

<p align="center">

<img

src="https://img.alicdn.com/imgextra/i1/O1CN01nTg6w21NqT5qFKH1u_!!6000000001621-55-tps-550-550.svg"

alt="AgentScope Logo"

width="200"

/>

</p>

<span align="center">

中文主页 | Documentation | Roadmap

</span>

<p align="center">

<a href="https://arxiv.org/abs/2402.14034">

<img

src="https://img.shields.io/badge/cs.MA-2402.14034-B31C1C?logo=arxiv&logoColor=B31C1C"

alt="arxiv"

/>

</a>

<a href="https://pypi.org/project/agentscope/">

<img

src="https://img.shields.io/badge/python-3.11+-blue?logo=python"

alt="pypi"

/>

</a>

<a href="https://pypi.org/project/agentscope/">

<img

src="https://img.shields.io/badge/dynamic/json?url=https%3A%2F%2Fpypi.org%2Fpypi%2Fagentscope%2Fjson&query=%24.info.version&prefix=v&logo=pypi&label=version"

alt="pypi"

/>

</a>

<a href="https://discord.gg/eYMpfnkG8h">

<img

src="https://img.shields.io/badge/Discord-Join%20Us-5865F2?logo=discord&logoColor=white"

alt="discord"

/>

</a>

<a href="https://docs.agentscope.io/">

<img

src="https://img.shields.io/badge/Docs-English%7C%E4%B8%AD%E6%96%87-blue?logo=markdown"

alt="docs"

/>

</a>

<a href="./LICENSE">

<img

src="https://img.shields.io/badge/license-Apache--2.0-black"

alt="license"

/>

</a>

</p>

<p align="center">

<img src="https://trendshift.io/api/badge/repositories/20310" alt="agentscope-ai%2Fagentscope | Trendshift" style="width: 250px; height: 55px;" width="250" height="55"/>

</p>

What is AgentScope 2.0?

AgentScope 2.0 is a production-ready, easy-to-use agent framework with essential abstractions that work with rising model capability and built-in support for .

• Event System → A unified event bus to the frontend and human-in-the-loop support.
• Permission System → Fine-grained, configurable control over tools and resources.
• Multi-tenancy & Multi-session Service → Production-grade serving with isolation across tenants and sessions.
• Workspace / Sandbox Support → Run tools and code in isolated environments, with built-in backends for local, Docker, E2B, OpenSandbox, and Daytona.
• Extensible Middleware System → Composable hooks to customize and extend the agent's reasoning-acting loop.

We design for increasingly agentic LLMs.

Our approach leverages the models' reasoning and tool use abilities

rather than constraining them with strict prompts and opinionated orchestrations.

<img src="assets/images/agentscope.png" alt="agentscope" width="100%"/>

News

<!-- BEGIN NEWS -->

• [2026-08] INTE: GitHub MCP Registry and ClawHub supported as built-in hubs. Example | Docs
• [2026-08] FEAT: MCP & Skill Hub supported — browse a hub, install into your library, add to a workspace. Example | Docs
• [2026-07] INTE: Daytona-based workspace/sandbox supported. Docs
• [2026-07] INTE: K8s, OpenSandbox-based workspace/sandbox supported. Docs
• [2026-07] INTE: ReMe long-term memory supported. Example | Docs
• [2026-06] FEAT: Agentic Memory supported. Example | Docs
• [2026-06] FEAT: Distributed & Multi-Tenancy & Multi-Session RAG service supported. Docs
• [2026-06] FEAT: RAG supported. Example | Docs
• [2026-06] INTE: Mem0 supported. Example | Docs
• [2026-06] FEAT: Agent Team supported. Example | Docs

<!-- END NEWS -->

More news →

Community

Welcome to join our community on

| Discord | DingTalk |

|----------------------------------------------------------------------------------------------------------------------------------|---------------------------------------------------------------------------|

| <img src="https://gw.alicdn.com/imgextra/i1/O1CN01hhD1mu1Dd3BWVUvxN_!!6000000000238-2-tps-400-400.png" width="100" height="100"> | <img src="./assets/images/dingtalk_qr_code.png" width="100" height="100"> |

Quickstart

Installation

AgentScope requires Python 3.11 or higher.

From PyPI

uv pip install agentscope
# or
# pip install agentscope

From source

# Pull the source code from GitHub
git clone -b main https://github.com/agentscope-ai/agentscope.git

# Install the package in editable mode
cd agentscope

uv pip install -e .
# or
# pip install -e .

Hello AgentScope!

Start your first agent with AgentScope 2.0:

from agentscope.agent import Agent
from agentscope.tool import Toolkit, Bash, Grep, Glob, Read, Write, Edit
from agentscope.credential import DashScopeCredential
from agentscope.model import DashScopeChatModel
from agentscope.message import UserMsg
from agentscope.event import EventType

import os, asyncio


async def main() -> None:
    agent = Agent(
        name="Friday",
        system_prompt="You're a helpful assistant named Friday.",
        model=DashScopeChatModel(
            credential=DashScopeCredential(
              api_key=os.environ["DASHSCOPE_API_KEY"]
            ),
            model="qwen3.6-plus",
        ),
        toolkit=Toolkit(
            tools=[
                Bash(),
                Grep(),
                Glob(),
                Read(),
                Write(),
                Edit(),
            ]
        ),
    )

    async for evt in agent.reply_stream(UserMsg("Tony", "Hi, Friday!")):
        # Handle the event stream, e.g., print the message, update UI, etc.
        match evt.type:
            case EventType.REPLY_START:
                ...
            case EventType.MODEL_CALL_START:
                ...
            case EventType.TEXT_BLOCK_START:
                ...
            case EventType.TEXT_BLOCK_DELTA:
                ...
            case EventType.TEXT_BLOCK_END:
                ...

            # Handle other event types

asyncio.run(main())

Hello Agent Service!

An extensible FastAPI based multi-tenancy, multi-session agent service with pre-built Web UI in examples/web_ui

<table>

<tr>

<td align="center">

<img src="assets/images/team.gif" alt="Agent team" width="100%"/>

<br/>

_{<b>Agent team</b> — a leader agent spawns workers and coordinates them through the built-in team tools.}

</td>

</tr>

<tr>

<td align="center">

<img src="assets/images/task.gif" alt="Task planning" width="100%"/>

<br/>

_{<b>Task planning</b> — the agent breaks complex work into a tracked plan and updates it as it goes.}

</td>

</tr>

<tr>

<td align="center">

<img src="assets/images/permission_bypass.gif" alt="Permission control in bypass mode" width="100%"/>

<br/>

_{<b>Permission control in bypass mode</b> — the agent runs end-to-end without pausing for tool-call confirmations.}

</td>

</tr>

<tr>

<td align="center">

<img src="assets/images/bg_tool.gif" alt="Background task offloading" width="100%"/>

<br/>

_{<b>Background task offloading</b> — a long-running tool moves to the background; its result later wakes the agent up and the conversation resumes.}

</td>

</tr>

</table>

Run the following commands to start the agent service backend and the web UI:

git clone -b main https://github.com/agentscope-ai/agentscope.git
cd agentscope/examples/agent_service

# start the agent service backend
python main.py

Then open another terminal to start the web UI:

cd agentscope/examples/web_ui

# start the webui
pnpm install
pnpm dev

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.

Publications

If you find our work helpful for your research or application, please cite our papers.

• AgentScope 1.0: A Developer-Centric Framework for Building Agentic Applications

• AgentScope: A Flexible yet Robust Multi-Agent Platform

@article{agentscope_v1,
    author  = {Dawei Gao, Zitao Li, Yuexiang Xie, Weirui Kuang, Liuyi Yao, Bingchen Qian, Zhijian Ma, Yue Cui, Haohao Luo, Shen Li, Lu Yi, Yi Yu, Shiqi He, Zhiling Luo, Wenmeng Zhou, Zhicheng Zhang, Xuguang He, Ziqian Chen, Weikai Liao, Farruh Isakulovich Kushnazarov, Yaliang Li, Bolin Ding, Jingren Zhou}
    title   = {AgentScope 1.0: A Developer-Centric Framework for Building Agentic Applications},
    journal = {CoRR},
    volume  = {abs/2508.16279},
    year    = {2025},
}

@article{agentscope,
    author  = {Dawei Gao, Zitao Li, Xuchen Pan, Weirui Kuang, Zhijian Ma, Bingchen Qian, Fei Wei, Wenhao Zhang, Yuexiang Xie, Daoyuan Chen, Liuyi Yao, Hongyi Peng, Zeyu Zhang, Lin Zhu, Chen Cheng, Hongzhu Shi, Yaliang Li, Bolin Ding, Jingren Zhou}
    title   = {AgentScope: A Flexible yet Robust Multi-Agent Platform},
    journal = {CoRR},
    volume  = {abs/2402.14034},
    year    = {2024},
}

Contributors

All thanks to our contributors:

<a href="https://github.com/agentscope-ai/agentscope/graphs/contributors">

<img src="https://contrib.rocks/image?repo=agentscope-ai/agentscope&max=999&columns=12&anon=1" />

</a>