Oralis

A living archive for endangered languages

Comment

Inspiration

I was binging a documentary on Youtube one time and realized that several languages have disappeared completely, taking entire cultures, histories, and identities with them.

The Akakari language became extinct in 2020 after the death of its last speaker, an elder woman named Ilicho. The Akabo dialect vanished in 2010 with Boa Senior, its final voice. The Jangil tribe was declared extinct as far back as 1920, leaving behind a language that was never even documented. Across the Great Andamanese peoples, once rich with over ten distinct dialects, most have been lost within a single century, just like that! No archive or repository to preserve these tribes' language or culture I built Oralis to ensure that no voice, no story, and no language has to disappear unnoticed again.

What it does

Oralis is a digital preservation platform that allows people to:

  • Record and upload audio contributions of endangered or native languages
  • Explore a global feed of linguistic repository of “memories” across cultures
  • Build a personal archive of their contributions
  • Discover languages through an intuitive, immersive interface

It transforms language preservation from a slow academic process into a living, participatory system where anyone can contribute to safeguarding cultural identity.

How I built it

Oralis is a full-stack system designed with scalability, security, and intentional UX in mind:

Frontend: Built with v0, delivering a clean, responsive, and immersive interface Backend APIs: Structured using serverless functions with strict validation and security layers Database: Powered by AWS DynamoDB with optimized GSIs for scalable querying Storage: AWS S3 with hardened presigned POST uploads, enforcing file size and type restrictions Pagination: Cursor-based infinite scrolling using DynamoDB’s LastEvaluatedKey Validation: Zod schemas ensure all inputs are sanitized and type-safe Security: Rate limiting to prevent abuse and unauthorized actions

Challenges we ran into

Designing a secure file upload system that prevents abuse while remaining seamless for users Implementing true pagination with DynamoDB without falling back to inefficient scans Ensuring authentication was enforced beyond the UI, especially at the API level Balancing performance and scalability in a serverless environment Creating a system that feels simple to use while handling complex backend logic These challenges enhanced my knowledge about the aws infrastructure.

Accomplishments that we're proud of

Building a secure, abuse-resistant upload pipeline using S3 presigned POST policies Implementing real-time scalable feeds with efficient database querying Designing a fully authenticated user ecosystem with ownership and data integrity Achieving a clean, cohesive UI/UX that reflects the underlying architecture Successfully compiling and deploying a system with zero build errors

Most importantly, we created a platform that turns passive awareness into active preservation.

What we learned

  • We learnt that scalability starts with data modeling, not just infrastructure
  • Security is not a just another feature
  • Serverless systems require thinking differently about state, limits, and reliability

What's next for Oralis

Introduce AI-powered transcription and translation for uploaded audio Expand into a language discovery map with geospatial visualization Enable community validation and moderation of contributions Add offline-first capabilities for remote and underserved regions Partner with linguists and cultural organizations to scale impact globally

My vision is to make Oralis the world’s living archive of human language before more cultures, languages and their tribal existence are lost forever.

Built With

Share this project:

Updates