# AgentScope

Build and run agents you can see, understand and trust.

Repository: https://github.com/agentscope-ai/agentscope
Canonical: https://ross.abutalabs.com/products/agentscope
Homepage: https://docs.agentscope.io/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: agent, chatbot, large-language-models, llm, llm-agent, multi-agent, multi-modal, mcp, react-agent
Last push: 2026-08-26T10:14:53+00:00
Link (homepage): https://docs.agentscope.io/
Link (site_page): https://docs.agentscope.io/versions/2.0.8dev/en/quickstart
Link (site_page): https://docs.agentscope.io/versions/2.0.8dev/en/others/faq
Link (site_page): https://docs.agentscope.io/versions/2.0.8dev/en/others/change-log

## Health v2 (maintenance only)
Score: 93/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 99, release rhythm 99, longevity 68
- inputs: {"age_days": 964, "days_push": 7, "days_rel": 9, "gap_med": 11.0, "n_releases_24m": 41}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 29706, forks 3409 (observed 2026-08-28T04:11:53.626249+00:00)

## What it is
AgentScope is a production-ready multi-agent framework for building LLM-powered agents with ReAct reasoning, tool execution, MCP integration, and human-in-the-loop oversight. Version 2.0 adds multi-tenant multi-session management, sandboxed workspaces (local, Docker, E2B, K8s), middleware, and a full-stack SDK with frontend UI and distributed deployment.

## Use cases
- build llm agents with tool calling
- create multi-agent systems in python
- run agents with human-in-the-loop approval
- sandbox agent tool execution in docker
- integrate mcp servers with agents
- deploy multi-tenant agent backend
- build chatbot with react reasoning
- stream agent events to a frontend

## When to choose
- you need a production-grade agent framework with security, permissions, and sandboxing
- you want multi-tenant, multi-session, or distributed agent deployment
- you need MCP, skills, and tool orchestration with human oversight
- you want a full-stack solution including SDK, UI, and backend

## When to avoid
- you rely on AgentScope 1.0 APIs, since 2.0 is a breaking release with no migration path
- you need built-in RAG and long-term memory today, as those modules are still being ported to 2.0
- you only need a lightweight single-prompt LLM call without agent machinery

## Facets
- artifact type: framework
- maturity: active
- function: agent-framework, llm-inference, mcp, chatbot, rag, middleware
- domain: large-language-models, artificial-intelligence, developer-tools
- platform: python, cross-platform, self-hosted
- tags: multi-agent, react-agent, human-in-the-loop, tool-orchestration, multi-tenant, distributed-agents, mcp-client, context-management, observability, sandboxing, ai-agents, docker, kubernetes

## Member repositories
- agentscope-ai/agentscope (main) score 93
- agentscope-ai/agentscope-java (sdk) score 83

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:11:53.626249+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T16:52:53.781414+00:00, confidence not recorded.
  - readme: https://github.com/agentscope-ai/agentscope (fetched 2026-08-28T04:11:53.626249+00:00, sha 5216107b171b)
  - homepage: https://docs.agentscope.io/ (fetched 2026-08-29T07:50:02.774673+00:00, sha 43bd19a22b37)
  - site_page: https://docs.agentscope.io/versions/2.0.8dev/en/quickstart (fetched 2026-08-29T07:50:02.784052+00:00, sha 0120c77fdbe5)
  - site_page: https://docs.agentscope.io/versions/2.0.8dev/en/others/faq (fetched 2026-08-29T07:50:02.790548+00:00, sha 10e6d621c070)
  - site_page: https://docs.agentscope.io/versions/2.0.8dev/en/others/change-log (fetched 2026-08-29T07:50:02.792639+00:00, sha 98d437b24bad)
  - registry_pypi: https://pypi.org/pypi/agentscope/json (fetched 2026-08-29T07:50:02.795132+00:00, sha 0210cc1c9ff4)
- Data as of 2026-08-30T08:39:29.467469+00:00.