Amazon Bedrock AgentCore Samples — Detailed Introduction
Amazon Bedrock AgentCore Samples is an official sample repository from AWS Labs that demonstrates how to build, deploy, and operate agentic AI applications using Amazon Bedrock AgentCore. The project is framework-agnostic and model-agnostic, providing reference implementations and step-by-step tutorials to accelerate productionization of AI agents.
Key components and content included:
-
Tutorials (01-tutorials): Notebook-based, hands-on guides that cover the major AgentCore components:
- Runtime: secure, serverless runtime examples for running agents at scale locally and in AWS.
- Gateway: examples showing how to convert APIs, Lambda functions, and services into tools usable by agents.
- Memory: managed memory examples for personalized, stateful agent experiences.
- Identity: integrations with identity providers (Okta, Entra, Cognito) and third-party apps (Slack, Zoom).
- Tools: built-in tools such as a Code Interpreter and Browser Tool with examples to extend agent capabilities.
- Observability: OpenTelemetry-compatible tracing and dashboards for debugging and monitoring agent workflows.
- End-to-end tutorials: moving a customer support agent from prototype to production.
-
Use cases (02-use-cases): End-to-end reference applications that show how to apply AgentCore features to real business scenarios.
-
Integrations (03-integrations): Examples integrating popular agent frameworks (Strands Agents, LangChain/LangGraph, CrewAI, LlamaIndex) and multi-agent communication patterns.
-
Infrastructure as Code (04-infrastructure-as-code): Deployment automation examples using CloudFormation, AWS CDK, and Terraform to provision production-ready resources.
-
Blueprints (05-blueprints): Full-stack, deployment-ready reference applications that include authentication, tooling, memory, and observability.
Highlights and distinguishing features:
- Framework & model agnostic: The samples show how to connect AgentCore to various agent frameworks and to any LLM available through Amazon Bedrock.
- Production focus: Emphasizes secure, scalable, and observable deployments with examples for runtime configuration, permissions, and CI/CD-friendly templates.
- Built-in enterprise tools: Demonstrates enterprise-grade Browser Tool and Code Interpreter tools that enable agents to interact with websites and execute code securely in sandboxed environments.
- Practical quickstarts: Quickstart guides and runnable notebook examples accelerate the path from local testing to full AWS deployment (agentcore configure / agentcore launch flows).
Getting started (summary):
- Clone the repository and inspect the notebook tutorials for the component you want to learn.
- Follow the Quick Start to set up a Python virtual environment, install dependencies, configure AWS credentials, and test locally.
- Use the provided IaC templates to deploy AgentCore runtime, gateways, tools, and observability in your AWS account.
This repository is intended for developers and engineering teams who want to productionize agentic AI solutions using AWS Bedrock AgentCore, providing both conceptual explanations and runnable reference implementations.
