Orchestrate
AI Agent Teams
Build, deploy and monitor collaborative AI agent teams. Let multiple Claude Code agents work together autonomously to solve complex tasks.
What is AgentCrew?
AgentCrew is an open-source platform that lets you create teams of AI agents powered by Claude Code. Each agent runs in its own container, has a specialized role, and collaborates through a message bus — enabling complex, autonomous workflows.
Multi-Agent Teams
Spin up entire teams of AI agents, each with a distinct role — leader, developer, reviewer, or any specialty you define.
Real-time Communication
Agents communicate through a NATS message bus. Watch live logs, WebSocket streams, and agent interactions in real time.
Any Infrastructure
Deploy on Docker for local development or scale to Kubernetes in production. Your infrastructure, your rules.
Everything you need to build
autonomous AI workflows
From team creation to production deployment — AgentCrew handles the orchestration so you can focus on results.
Create Teams
Configure teams of specialized agents with custom roles, instructions, and permissions through an intuitive wizard.
Deploy Instantly
Launch agent containers on Docker or Kubernetes with a single click. Infrastructure is provisioned automatically.
Monitor in Real-Time
Watch your agents work via live logs, WebSocket streams, and a real-time chat interface.
Leader + Workers
Define a leader agent that coordinates tasks and worker agents that execute them in parallel.
How it works
From zero to a running AI agent team in minutes. No complex setup required.
Create your team
Name your team, add agents, and define each agent's role (leader or worker) through the team builder.
Configure agents
Write a CLAUDE.md for each agent with their specialty, instructions, and behavior. Fully customizable.
Deploy
Click deploy. AgentCrew provisions a Docker network, NATS message bus, and starts all agent containers automatically.
Chat & Monitor
Send tasks to your team through the chat interface. Watch agents collaborate in real-time via live logs and activity streams.
Built with modern tech
Production-ready technologies chosen for performance, reliability, and developer experience.
System Architecture
A clean, event-driven architecture where agents communicate asynchronously through a central message bus.
Isolated Containers
Each agent runs in its own container with isolated file system and tools.
Full Observability
Stream logs, agent states, and task progress via WebSocket in real time.
Async by Design
NATS enables non-blocking, pub/sub communication between all agents.
Start building your
agent team today
Open source, self-hosted, and ready to run in minutes.