Your first function with AWS SAM

AWS SAM (Serverless Application Model) is an extension of CloudFormation with shorthand for serverless resources, plus a CLI for building, testing locally and deploying.

Create the project

sam init --runtime python3.13 --app-template hello-world --name bookings-api
cd bookings-api

The generated template.yaml is the heart of the project. A trimmed version:

AWSTemplateFormatVersion: "2010-09-09"
Transform: AWS::Serverless-2016-10-31

Globals:
  Function:
    Runtime: python3.13
    Architectures: [arm64]
    MemorySize: 256
    Timeout: 10

Resources:
  ListSessionsFunction:
    Type: AWS::Serverless::Function
    Properties:
      CodeUri: src/
      Handler: sessions.list_handler
      Events:
        Api:
          Type: HttpApi
          Properties:
            Path: /sessions
            Method: get

The Transform line tells CloudFormation to expand SAM resources into regular ones. One AWS::Serverless::Function with an HttpApi event creates the function, its execution role, an HTTP API, a route and the permission for API Gateway to invoke the function.

Why arm64

Graviton based Lambda functions are priced lower per gigabyte second than x86 and perform well for most Python and Node.js code. Unless a dependency ships only x86 binaries, choose arm64.

The handler

# src/sessions.py
import json

def list_handler(event, context):
    sessions = [{"id": "s-1", "title": "Intro to pottery", "seats_left": 4}]
    return {
        "statusCode": 200,
        "headers": {"content-type": "application/json"},
        "body": json.dumps(sessions),
    }

Build, run locally, deploy

sam build
sam local start-api          # runs the API on localhost using Docker
curl http://127.0.0.1:3000/sessions

sam deploy --guided          # first deploy asks for stack name and region

sam deploy --guided saves your answers to samconfig.toml, so later deploys are just sam deploy. At the end, SAM prints the API endpoint URL. Call it with curl and you have a live serverless API.

Read what was created

Open CloudFormation in the console and look at the stack's resources. Knowing exactly which role, API and permission SAM generated is the first step to controlling them, which we do in the next lessons.