VRSEN/agency-swarm

Reliable Multi-Agent Orchestration Framework

What it solves

Agency Swarm simplifies the creation of multi-agent AI applications by organizing automation around real-world organizational structures. It provides a structured way to define agent roles, manage their communication, and deploy them as a collaborative "swarm" to handle complex tasks.

How it works

Built on top of the OpenAI Agents SDK, the framework allows users to define individual Agent objects with specific roles (e.g., CEO, Developer), instructions, and tools. These agents are then organized into an Agency where directional communication_flows (using the > operator) dictate which agents can initiate contact with others. The framework uses Pydantic models for type-safe tool development and supports state persistence via custom callbacks for saving and loading conversation threads.

Who it’s for

Developers building production-ready multi-agent systems who want an intuitive, organizational-based approach to agent orchestration and a structured folder system for managing agent assets.

Highlights

  • Organizational Orchestration: Uses a directional flow system to manage how agents communicate and collaborate.
  • Type-Safe Tooling: Supports Pydantic-based tools and the ability to convert OpenAPI schemas directly into agent tools.
  • Flexible Model Support: Works natively with OpenAI models and supports others (Claude, Gemini, Grok) via LiteLLM.
  • Built-in Demos: Includes a Copilot-style web UI and a Terminal User Interface (TUI) for rapid testing and the execution of agent workflows.

Related

  • Project
    kyegomez/swarms

    An enterprise-grade multi-agent orchestration framework that provides pre-built architectures for coordinating multiple AI agents in sequential, concurrent, and hierarchical workflows.

  • Project
    desplega-ai/agent-swarm

    An open-source AI operating system that coordinates a lead agent and Docker-isolated worker agents to automate complex workflows across Slack, GitHub, and email.

  • Project
    wanikua/danghuangshang

    A multi-agent collaboration system based on the OpenClaw framework that organizes AI agents into a hierarchical government or corporate structure for efficient task delegation and execution.

  • Project
    2FastLabs/agent-squad

    A flexible, lightweight framework for orchestrating multiple AI agents across Python, TypeScript, and Swift, featuring intelligent routing and context management.

  • Project
    openswarm-ai/openswarm

    A local orchestrator for managing multiple AI agents in parallel via a spatial dashboard, featuring Git worktree isolation and unified tool approvals.