Project
Text-to-Speech Serverless App
Serverless multilingual text-to-speech app on AWS — Cognito auth, Lambda processing pipeline, Polly synthesis, and Terraform IaC.
- Type
- Cloud / Serverless
- Status
- Completed
- Year
- AWS
- Lambda
- Polly
- Cognito
- DynamoDB
- Terraform
- API Gateway
- SNS
- S3
- React
What I Built
A full-stack serverless application that translates text into any supported language and synthesizes it as spoken audio — secured with Cognito, processed asynchronously, and deployed entirely with Terraform.
- React frontend hosted on Amplify with Cognito sign-up/sign-in and JWT-based API authorization.
- Submission pipeline — API Gateway receives text, Lambda writes to DynamoDB and publishes to SNS for async processing.
- Processing pipeline — A second Lambda translates the text via Amazon Translate, synthesizes speech with Polly, and stores the audio in S3.
- Retrieval — The frontend polls a third Lambda to fetch the completed audio file URL from DynamoDB.
- Infrastructure as Code — Every resource defined in Terraform.
Architecture
User → React (Amplify) → Cognito Auth → API Gateway → Lambda (submit)
↓
DynamoDB + SNS
↓
Lambda (process: Translate + Polly)
↓
S3 (audio) + DynamoDB (update)
↓
Lambda (retrieve) → React UI