Skip to main content
Bedrock Chat offers multiple deployment methods to suit different use cases and environments. All deployment methods use AWS CDK internally to provision infrastructure.

​
Deployment Methods

Bedrock Chat provides three primary deployment approaches: The easiest way to deploy Bedrock Chat using AWS CloudShell. This method uses the bin.sh script and AWS CodeBuild to handle the deployment automatically.
  • Best for: Quick deployments, testing, and production use
  • Time: ~35 minutes
  • Prerequisites: AWS account with Bedrock access
  • Automation: Fully automated via CodeBuild
View CloudShell deployment guide →

​
Direct CDK Deployment

Deploy directly using AWS CDK from your local machine. This method gives you more control and is ideal for development.
  • Best for: Development, customization, and CI/CD pipelines
  • Time: Variable (depends on your environment)
  • Prerequisites: UNIX/Linux, Docker, Node.js runtime
  • Control: Full control over the deployment process
View CDK deployment guide →

​
Multi-Environment Deployment

Deploy multiple environments (dev, staging, prod) from the same codebase using the parameter.ts configuration file.
  • Best for: Organizations managing multiple environments
  • Configuration: Type-safe parameter management
  • Isolation: Separate resources per environment
View multi-environment guide →

​
Architecture

Bedrock Chat is built on AWS managed services:
  • Amazon DynamoDB: Conversation history storage
  • Amazon API Gateway + AWS Lambda: Backend API (FastAPI with Lambda Web Adapter)
  • Amazon CloudFront + S3: Frontend delivery (React, Tailwind CSS)
  • AWS WAF: IP address restrictions
  • Amazon Cognito: User authentication
  • Amazon Bedrock: Foundation models via API
  • Amazon Bedrock Knowledge Bases: RAG capabilities
  • Amazon OpenSearch Serverless: Vector search backend
  • Amazon EventBridge Pipes: Bot deletion events
  • AWS Step Functions: Knowledge base ingestion orchestration
  • Amazon Athena: Usage analytics

​
Prerequisites

Before deploying Bedrock Chat, ensure you have:

​
AWS Account Requirements

  1. Bedrock Model Access: Enable model access in the us-east-1 region
    • Go to Bedrock Model access
    • Click Manage model access
    • Select the models you want to use
    • Click Save changes
  2. Region Selection: Deploy in a region where OpenSearch Serverless is available
  3. IAM Permissions: Administrator access or equivalent permissions to create:
    • CloudFormation stacks
    • S3 buckets
    • Lambda functions
    • CloudFront distributions
    • Cognito user pools
    • OpenSearch Serverless collections
    • IAM roles and policies

​
Security Considerations

By default, Bedrock Chat allows anyone who knows the URL to sign up. For production deployments, you should:
  • Add IP address restrictions using --ipv4-ranges and --ipv6-ranges
  • Disable self-registration with --disable-self-register
  • Restrict signup email domains with --allowed-signup-email-domains

​
Deployment Outputs

After successful deployment, you’ll receive:
  • Frontend URL: CloudFront distribution URL (e.g., https://xxxxxxxxx.cloudfront.net)
  • User Pool ID: Cognito user pool for managing users
  • User Pool Client ID: Cognito app client ID
  • Backend API URL: API Gateway endpoint

​
Next Steps

CloudShell Deployment

Deploy quickly using AWS CloudShell

CDK Deployment

Deploy using AWS CDK directly

Supported Regions

Check region availability and requirements

Parameters Reference

Explore all deployment parameters

Build docs developers (and LLMs) love

Get started for freeTalk to us