Terraform Apply

This page walks through the complete deployment process — from cloning the source repo to verifying your running application. The entire stack deploys in a single Terraform Terraform
HashiCorp Terraform — infrastructure-as-code tool used to provision all AWS resources in this demo. apply that takes around 3–5 minutes.

Clone the source repo

Start by cloning the infrastructure repository:

Terminal

git clone https://github.com/jajera/terraform-aws-s3-vectors-rag-demo.git
cd terraform-aws-s3-vectors-rag-demo

Configure terraform.tfvars

Copy the example variables file and edit it with your values:

Terminal

cp terraform.tfvars.example terraform.tfvars

Open terraform.tfvars in your editor and set the required variables:

terraform.tfvars

project              = "s3vec-rag"
environment          = "dev"
aws_region           = "us-east-1"
cognito_admin_email  = "you@example.com"

rss_feeds = [
  "https://aws.amazon.com/about-aws/whats-new/recent-announcements/feed/",
]

Note

The cognito_admin_email receives a temporary password after deployment. Use a real email address you can access.

Variable reference

Variable	Required	Description
project	Yes	Name prefix for all AWS resources (keep short, lowercase, no spaces)
environment	Yes	Deployment environment label (e.g., dev, staging, prod)
aws_region	Yes	AWS region — must support both S3 Vectors S3 Vectors Amazon S3 Vectors — a purpose-built vector storage capability within S3 that enables similarity search over embeddings without a separate vector database. and Bedrock Bedrock Amazon Bedrock — a fully managed service for accessing foundation models (embedding and LLM) via a unified API.
cognito_admin_email	Yes	Email for the auto-created admin user in Cognito Cognito Amazon Cognito — a user authentication service providing user pools, JWT tokens, and group-based authorization.
rss_feeds	Yes	List of RSS feed URLs to ingest into the vector store
ingest_schedule_expression	No	EventBridge EventBridge Amazon EventBridge — a serverless event bus. Used here for the daily scheduled corpus ingest cron trigger. schedule expression (default: cron(0 6 * * ? *) — daily at 06:00 UTC)
embedding_model_id	No	Embedding model Embedding model Amazon Titan Embeddings V2 (amazon.titan-embed-text-v2:0) — converts text into 1024-dimensional vectors for similarity search. ID (default: amazon.titan-embed-text-v2:0)
inference_profile_id	No	Bedrock inference profile for the LLM (default: Claude via cross-region profile)

Caution

Choose a region that supports both S3 Vectors and the Bedrock models you need. As of mid-2025, us-east-1 and us-west-2 are recommended.

Initialize Terraform

Download the required providers and modules:

Terminal

terraform init

You should see output confirming the AWS provider was installed:

Initializing provider plugins...
- Installing hashicorp/aws v6.x.x...
- Installing hashicorp/random v3.x.x...
- Installing hashicorp/null v3.x.x...

Terraform has been successfully initialized!

Preview the plan

Before applying, review what Terraform will create:

Terminal

terraform plan

This shows the full list of resources to be provisioned without making any changes. Expect around 40–50 resources including Lambda Lambda
AWS Lambda — serverless compute. This project uses two functions: ingest (RSS → embed → store) and query (search → answer). functions, API Gateway API Gateway
Amazon API Gateway — a managed HTTP API service with JWT authorization, CORS, and throttling. Routes requests to the Query Lambda. routes, IAM roles, and the Amplify Amplify
AWS Amplify — a managed hosting service for static web applications. Used to host the briefing UI SPA. app.

Tip

Review the plan output to confirm your region and project name are correct before applying.

Apply the configuration

Deploy the entire stack:

Terminal

terraform apply

Type yes when prompted to confirm. The deployment takes approximately 3–5 minutes.

What happens during apply

Terraform provisions resources in this sequence:

1. Foundation — IAM roles, S3 buckets, vector bucket vector bucket
   A specialized S3 bucket type (aws_s3vectors_vector_bucket) that hosts vector indexes for similarity search. , and vector index vector index
   A named index (aws_s3vectors_index) within a vector bucket defining dimension, distance metric, and data type.
2. Compute — Lambda functions for ingest and query, with Python dependencies packaged as layers
3. API layer — API Gateway HTTP API with JWT authorizer and CORS configuration
4. Auth — Cognito user pool, app client, admin user, and user groups
5. Frontend — Amplify app with the React SPA built and deployed
6. Scheduling — EventBridge rule for daily ingest cron
7. Bootstrap — A null_resource provisioner triggers the initial ingest, embedding ~120 articles into the vector store

Note

The bootstrap ingest runs automatically as part of the apply. You will have searchable content immediately after deployment completes.

Verify outputs

Once the apply finishes, retrieve the key outputs:

Terminal

terraform output

This prints all configured outputs. To get individual values:

Terminal

terraform output -raw app_url
terraform output -raw api_endpoint
terraform output -raw cognito_user_pool_id

Output reference

Output	Description
app_url	Amplify HTTPS URL for the web UI — open this in your browser
api_endpoint	API Gateway base URL for the query and ingest endpoints
cognito_user_pool_id	Cognito user pool ID (needed for CLI user management)
cognito_client_id	SPA client ID used by the frontend for authentication
vector_bucket_name	The S3 Vectors bucket name (useful for debugging)
ingest_function_name	Ingest Lambda function name (for manual invocations or log inspection)

Tip

Copy the app_url and open it in your browser. You will need the Cognito credentials (sent to your admin email) to log in. See the Cognito Users page for details.

Next steps

• Set up additional users — create readers and manage groups
• Explore the web UI — understand the frontend and redeployment
• Tear down the stack — clean up all resources when done