Welcome to the Amazon Bedrock AgentCore Samples repository!
Amazon Bedrock AgentCore is both framework-agnostic and model-agnostic, giving you the flexibility to deploy and operate advanced AI agents securely and at scale. Whether you’re building with Strands Agents, CrewAI, LangGraph, LlamaIndex, or any other framework—and running them on any Large Language Model (LLM)—Amazon Bedrock AgentCore provides the infrastructure to support them. By eliminating the undifferentiated heavy lifting of building and managing specialized agent infrastructure, Amazon Bedrock AgentCore lets you bring your preferred framework and model, and deploy without rewriting code.
This collection provides examples and tutorials to help you understand, implement, and integrate Amazon Bedrock AgentCore capabilities into your applications.
Migrating from the Starter Toolkit? This repository is transitioning from the Bedrock AgentCore Starter Toolkit to the new AgentCore CLI. Samples that still depend on the Starter Toolkit are in legacy/ and will be updated over the coming weeks. See MIGRATION.md for the full old-path to new-path mapping.
🎥 Video
Build your first production-ready AI agent with Amazon Bedrock AgentCore. We’ll take you beyond prototyping and show you how to productionize your first agentic AI application using Amazon Bedrock AgentCore.
📁 Repository Structure
Your First Agent in Minutes
Get up and running with the AgentCore CLI — the fastest way to create, develop, and deploy agents on Amazon Bedrock AgentCore.
python/ — Python agent samples (Code Interpreter, Gateway, Memory, Identity, and more)typescript/ — TypeScript agent samples
AgentCore Capabilities Deep Dives
Focused examples for individual AgentCore capabilities:
- Runtime — Secure, serverless runtime for deploying agents and tools at scale
- Gateway — Convert APIs, Lambda functions, and services into MCP-compatible tools
- Identity — Agent identity and access management across AWS and third-party apps
- Memory — Managed memory infrastructure for personalized agent experiences
- Tools — Built-in Code Interpreter, Browser Tool, Web Search Tool
- Observability — Trace, debug, and monitor agent performance with OpenTelemetry
- Evaluation — Built-in and custom evaluators for on-demand and online evaluation
- Policy — Fine-grained access control with Cedar policies
Complete Applications
Production-ready use cases that combine multiple AgentCore capabilities to solve real business problems. Each includes deployment instructions, architecture diagrams, and testing guides.
Connect AgentCore to Your Stack
identity-providers/ — Okta, Entra, Cognito, and other IdP integrationsobservability/ — Grafana, Datadog, Dynatrace, and other monitoring platformsdata-platforms/ — Data lake, warehouse, and analytics integrationsux-examples/ — Streamlit, AG-UI, and other frontend patterns
Deployment Automation
Production-ready templates for provisioning AgentCore resources with CloudFormation, AWS CDK, or Terraform.
Full-Stack Reference Applications
Complete, deployment-ready agentic applications with integrated services, authentication, and business logic you can customize for your use case.
Quick Start with the AgentCore CLI
The AgentCore CLI is the recommended way to create, develop, and deploy agents on Amazon Bedrock AgentCore. It replaces the previous Starter Toolkit with a streamlined project-based workflow.
Step 1: Prerequisites
- An AWS account with credentials configured (
aws configure)
- Node.js 20.x or later
uv (for Python agents) or Node.js (for TypeScript agents)- Model Access: Anthropic Claude 4.0 enabled in Amazon Bedrock console
- AWS Permissions:
BedrockAgentCoreFullAccess managed policyAmazonBedrockFullAccess managed policy
Step 2: Install the CLI and Create a Project
# Install the AgentCore CLI
npm install -g @aws/agentcore
# Create a new project (interactive wizard)
agentcore create
cd my-agent
The create wizard scaffolds a ready-to-run project with your choice of framework (Strands Agents, LangGraph, Google ADK, OpenAI, and more) and language (Python or TypeScript).
This is the beginning of the README. Read the rest at the original. Read the full README ↗
