VRSEN/agency-swarm

Health v2 · maintenance only

Adoption not part of the score

4540 stars · 1056 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

Agency Swarm is a Python framework for building multi-agent AI applications on top of the OpenAI Agents SDK. It lets developers define specialized agent roles, orchestrate communication flows between them, and deploy agencies as FastAPI or MCP endpoints.

Use cases

• build a multi-agent AI system with distinct roles like CEO and assistant
• orchestrate collaborative AI agents that talk to each other
• serve AI agent tools as MCP endpoints
• deploy a team of LLM agents as a production API
• create custom AI agents with type-safe Pydantic tools
• persist multi-agent conversation threads in a database
• run agents on Anthropic, Gemini, or local models via LiteLLM

When to choose

• you want structured, role-based multi-agent orchestration with explicit communication flows
• you need production deployment features like FastAPI serving, auth, and thread persistence
• you want full control over agent prompts and type-safe tool definitions
• you need MCP or third-party model support (Azure, Anthropic, Ollama)

When to avoid

• you need a single-agent chatbot with no orchestration
• you want a provider-agnostic framework not tied to the OpenAI Agents SDK
• you need a lightweight agent loop without role/agency abstractions
• your project is not in Python

Facets

framework · maturity active

agent-framework llm-inference mcp api-framework workflow-automation artificial-intelligence large-language-models developer-tools python cross-platform self-hosted multi-agent orchestration openai-agents-sdk swarm fastapi litellm pydantic production-ready ai-agents automation

Member repositories

For agents

markdown · JSON · MCP: product_card(name="VRSEN/agency-swarm")