Infrastructure Architecture

Compact reference for Digital Thai MoCA / BrainDI platform infrastructure.

AWS account: 564141170168 · Region: ap-southeast-1

Repositories & Deploy Targets

Repository	Deploy target	Runtime
automoca-web-app	AWS Amplify → CloudFront	Static SPA
braindi-backoffice-frontend	AWS Amplify → CloudFront	Static SPA
automoca-api-services	ECS (braindi-backend) + ALB	Go API (port 8080)
automoca-aiservice	ECS (automoca-ai-service-dev)	Python Kafka worker

automoca-api-services is deployed in AWS as braindi-backend (ECR, ECS, secrets).

Topology

DNS (Route53 → CloudFront / ALB)

Environment	Patient app	Backoffice	API
Prod DTM	dtm.aimet.tech	backoffice.dtm.aimet.tech	api.dtm.aimet.tech
Prod BrainDI	braindi.aimet.tech	backoffice.braindi.aimet.tech	api.braindi.aimet.tech
Dev	dev.dtm.aimet.tech	backoffice.dev.dtm.aimet.tech	api.dev.braindi.aimet.tech
Siriraj	siriraj.*.aimet.tech	backoffice.siriraj.*.aimet.tech	api.siriraj.*.aimet.tech

• Frontends: CloudFront distributions
• Prod API: braindi-prod-1392139716.ap-southeast-1.elb.amazonaws.com
• Dev API: dev-alb-452655975.ap-southeast-1.elb.amazonaws.com

ECS

Cluster	Service	Task definition
braindi-backend	braindi-backend-dev	braindi-backend-dev
braindi-backend	braindi-backend-prod	braindi-backend-prod
automoca-ai-service-dev	automoca-ai-service-dev-service-*	automoca-ai-service-dev
automoca-ai-service-dev	automoca-ai-service-prod-service-*	automoca-ai-service-prod

Global ASR Integration

Shared platform used by braindi-backend (not deployed from this repo).

Mode	Path	Use case
Real-time	Patient app → WSS /web/tests/ws/speech → backend WebSocket reverse proxy → ASR_WS_URL/ws/v1/speech	Live transcription during MoCA test (via @bream-is-a-fish/aimet-asr-react-client)
Async	Backend → Kafka input.asr → global-asr-service → Kafka braindi.output.asr → backend consumer	Batch/offline audio transcription

Backend env: ASR_WS_URL (WebSocket base URL), ASR_KAFKA_* (async topics/brokers).

Global ASR Service ECS

Cluster	Service	Task definition
global-asr-service	global-asr-api-server-dev	global-asr-api-server-dev
global-asr-service	global-asr-consumer-dev	global-asr-consumer-dev
global-asr-service	global-asr-api-server-prod	global-asr-api-server-prod
global-asr-service	global-asr-consumer-prod	global-asr-consumer-prod

• api-server — real-time WebSocket (/ws/v1/speech)
• consumer — async Kafka (input.asr → braindi.output.asr)

Global ASR Service S3

Bucket	Notes
global-asr-service-dev	Async ASR input storage (dev). Backend uploads audio for Kafka input.asr jobs; consumer reads and transcribes.
global-asr-service-prod	Same as dev (prod).

Kafka

• Topics: input.asr → braindi.output.asr
• IAM auth in prod

ECR

Repository	Image example
braindi-backend	564141170168.dkr.ecr.ap-southeast-1.amazonaws.com/braindi-backend:<tag>
automoca-ai-service	564141170168.dkr.ecr.ap-southeast-1.amazonaws.com/automoca-ai-service:<sha>

Amplify

App	Repository	Default domain
auto-moca-web-app	automoca-web-app	ddvcmytiooo28.amplifyapp.com
braindi-backoffice-frontend	braindi-backoffice-frontend	d36kp5f38yecvq.amplifyapp.com

Build: yarn install → yarn build → artifacts from dist/. Sentry secrets injected from AWS Secrets Manager at build time.

Data Stores

Service	Resource	Notes
MongoDB	External (Atlas)	MONGO_URI via Secrets Manager
Redis	braindi-backend-prod-0001-001	ElastiCache, cache.t4g.micro
S3	automoca-web-backend-dev	Primary storage: test uploads (audio/images), generated PDFs. Used by braindi-backend-dev, automoca-ai-service-dev. Keys: {testId}/{uuid}_{filename}
S3	automoca-web-backend-prod	Same as dev. Used by braindi-backend-prod, automoca-ai-service-prod

Messaging

Component	Topics / channels
Kafka (IAM auth in prod)	AI: input.ai.braindi → output.ai.braindi
SNS	braindi-predict-result-delay-alarm-dev/prod
automoca-aiservice (ECS)	Consumes input.ai.braindi, produces output.ai.braindi

Secrets Manager

Secret prefix	Used by
braindi-backend-dev, braindi-backend-prod	Go API
dev/prod/automoca/ai-service/*	AI service (GCP, Azure, API keys)

CI/CD

Repo	Trigger	Pipeline
automoca-api-services	Push to dev	GitHub Actions → ECR → ECS braindi-backend-dev
automoca-aiservice	Push to dev / main	GitHub Actions → ECR → ECS dev / update prod task def
automoca-web-app	Git push	Amplify auto-deploy
braindi-backoffice-frontend	Git push	Amplify auto-deploy

Backend deploy role: arn:aws:iam::564141170168:role/braindi-backend-actions-runner

External Dependencies (outside this AWS account)

Service	Purpose
Google Cloud Speech-to-Text	Speech recognition
Google Gemini / Azure OpenAI	AI scoring (via aiservice)
CFC proxy (VITE_CFC_PROXY_ENDPOINT)	Backoffice file upload
Sentry	Frontend error tracking