strands-agents/samples

Strands Agents Samples – Quick Overview

What it is: A collection of ready‑made example projects that show how to build AI agents with the Strands platform. The repo contains step‑by‑step tutorials for both Python and TypeScript, covering everything from a tiny “hello‑world” agent to more complex patterns like multi‑agent systems, RAG (retrieval‑augmented generation), deployment to cloud services, and edge‑device integration.

Why it matters: Strands promotes a model‑driven way of defining agents – you describe the agent’s behavior (system prompt, tools, routing) in a concise configuration, and the SDK handles the heavy lifting (LLM calls, streaming, tool execution). These samples let developers see that workflow in action without writing a lot of boilerplate.

Key parts of the repo:

• Python samples (python/):
  • 01‑learn – basic SDK usage, creating an Agent, streaming responses.
  • 02‑deploy – how to package an agent for AWS Lambda, Fargate, or the Strands AgentCore runtime.
  • 03‑integrate – connecting agents to AWS services, databases, or third‑party APIs.
  • 04‑industry‑use‑cases – example prompts and architectures for finance, healthcare, retail, etc.
  • 05‑technical‑use‑cases – patterns such as Agentic RAG, tool calling, and orchestration.
  • 06‑evaluate – testing and evaluation utilities.
  • 07‑ux‑demos – full‑stack UI demos that talk to an agent.
  • 08‑edge – running agents on edge devices, including robotics.
• TypeScript samples (typescript/):
  • Mirrors the Python tutorials but for the Node.js SDK (@strands‑agents/sdk).
  • Includes deployment guidance for the Strands AgentCore runtime.

How to get started:

1. Install the SDK (pip install strands-agents strands-agents-tools or npm install @strands-agents/sdk).
2. Follow the “Your First Agent” snippet in the README – a few lines create an Agent and invoke it with a prompt.
3. Configure a model provider (OpenAI, Anthropic, etc.) as described in the Strands docs.
4. Dive into the numbered tutorial folders for deeper examples.

Intended audience: Developers who want a hands‑on introduction to building LLM‑backed agents, especially those interested in a declarative, model‑centric approach. The repo is educational; it warns that the code is not production‑ready without additional security and testing.

License & community: Apache‑2.0, with a public Discord for support and a contribution guide for pull‑requests.

---

All details are taken directly from the repository’s README.