awslabs/amazon-bedrock-agent-samples

What it solves

This repository provides practical examples and best practices for building AI-powered assistants using Amazon Bedrock Agents. It helps developers move from experimentation to production by demonstrating how to automate complex workflows and interact with various APIs and services.

How it works

The project uses the Amazon Bedrock Agents framework to create agents that can execute tasks. It supports both single-agent setups and multi-agent collaboration, where a supervisor can plan and execute tasks across multiple agents using routing or supervisor mode. The examples are primarily implemented using the boto3 SDK in Python, but they are compatible with various other AWS SDKs.

Who it’s for

Developers and architects who are using Amazon Bedrock to build agentic AI solutions, specifically those looking for reference implementations of common agent patterns (e.g., code interpretation, long-term memory, human-in-the-loop).

Highlights

• Multi-Agent Collaboration: Support for supervisor-led orchestration of multiple agents for complex problem solving.
• Diverse Example Library: Includes implementations for a wide range of use cases like DevOps, energy efficiency, mortgage assistance, and real estate investment.
• Advanced Agent Capabilities: Examples covering code interpretation, knowledge base integration, long-term memory, and user confirmation steps.
• Integrated Tooling: Shared tools for web search, working memory, and stock data lookup that can be reused via Action Groups.