Deploy CrewAI multi-agent teams with Amazon Bedrock Agent Core
Local development → production-ready serverless deployment. No infrastructure management.

Quick Start · Structure · Walkthrough · Prerequisites

---

From local development to production-ready serverless deployment — no infrastructure management required. This repo contains the complete, validated code from Blog #8: Deploying CrewAI Multi-Agent Teams with Amazon Bedrock Agent Core.

The core idea: You build your AI agents locally with CrewAI, wrap them with a single @app.entrypoint decorator, and Agent Core handles the rest — scaling, security, session isolation, and observability.

What You'll Learn

• CrewAI Multi-Agent Setup — Research, Analysis, and Writing agents collaborating on tasks
• Bedrock Integration — Using Claude models through CrewAI's LLM wrapper
• Agent Core Deployment — The full create → dev → deploy → invoke workflow
• Production Patterns — Error handling, rate limiting, and real-world deployment considerations

Repo Structure

├── bedrock-agentcore-crewai-validation.ipynb  # 📓 Full walkthrough notebook (start here)
├── crewai_agent.py                            # Standalone CrewAI agent for local testing
├── setup_env.sh                               # Environment setup script
├── pyproject.toml                             # Python dependencies
└── CrewAIAgent/                               # Agent Core project (generated + customized)
    ├── agentcore/                             # Infrastructure config & CDK
    └── app/CrewAIAgent/                       # Production agent code
        ├── main.py                            # Entry point with @app.entrypoint
        └── pyproject.toml                     # Agent dependencies

Quick Start

# 1. Clone and setup
git clone https://github.com/breakingthecloud/crewai-agentcore-deployment.git
cd crewai-agentcore-deployment
chmod +x setup_env.sh && ./setup_env.sh

# 2. Activate virtual environment
source .venv/bin/activate

# 3. Test locally
python crewai_agent.py

# 4. Open the notebook for the full walkthrough
jupyter lab bedrock-agentcore-crewai-validation.ipynb

Notebook Walkthrough

Section	Description
1. Setup	Environment configuration and dependency installation
2. Local CrewAI	Building a multi-agent research team locally
3. Agent Core Project	Creating and configuring the Agent Core project
4. Local Dev Testing	Testing the agent with agent-core dev
5. Production Deployment	Deploying with agent-core deploy
6. Invocation	Testing the deployed agent via agent-core invoke
7. Production Patterns	Error handling, rate limits, real-world considerations

Prerequisites

• AWS account with Bedrock model access enabled (Claude Sonnet)
• Python 3.10+
• Node.js 20+ (for Agent Core CLI)
• AWS credentials configured (aws configure or SSO)

Key Takeaway

Agent Core is framework-agnostic — it works with CrewAI, LangGraph, Strands Agents, Google ADK, OpenAI Agents, or any custom framework. The deployment pattern is always the same: wrap your agent logic with BedrockAgentCoreApp and let AWS handle production infrastructure.

---

📝 Blog post: breakingthecloud.com

License

Apache 2.0.

---

breakingthecloud.com · cortez.cloud

_{Build locally. Deploy serverlessly. Scale infinitely.}