Get Started with AgentCore
This page helps you pick the right getting-started path. Whether you're a first-time agent builder evaluating managed vs. code-first approaches, or an experienced developer migrating an existing agent onto AgentCore Runtime, the decision tree below will route you to the right tutorial.
Pick your path
| If you want to… | Start here |
|---|---|
| Deploy an agent through the console without writing orchestration code | Your First Agent (Console) |
| Build a code-first prototype from your terminal | Build from CLI |
| Migrate an existing agent (Strands, LangGraph, CrewAI, custom) onto AgentCore | Existing Agent |
| Compare all deployment methods before deciding | Deployment Methods |
Start with the Harness getting-started guide for the fastest route to a running agent. Switch to the CLI Python tutorial if you prefer writing agent logic in code.
Before you begin
Every path requires the same account and environment preparation. Complete the Preflight Checklist before starting any tutorial — it covers region selection, IAM permissions, model access, and cost awareness.
What each path produces
Each Get Started tutorial produces a working prototype: a deployed agent with a reachable endpoint you can invoke and iterate on. What remains after that is applying production controls — security, observability, policy enforcement, and operational readiness.
Move to Preparing for production once you have a working prototype and want to serve real user traffic.
Everything produced by a Get Started tutorial is a prototype until you have applied the controls in Preparing for production. Do not use a prototype endpoint to serve real user traffic.