RagamNesia

Inspiration## Inspiration

We are inspired to create RagamNesia because we believe in the importance of keeping Indonesia united in its diversity. In a world that's rapidly changing and becoming more globalized, we want to ensure that the unique richness of our local languages, dialects, and cultural traditions are not lost or forgotten. Our goal is to preserve and celebrate the vibrant cultural identity of Indonesia — not just as a memory of the past, but as a living, evolving part of our future. By empowering the younger generation to learn and appreciate their heritage, we hope to strengthen the spirit of Bhinneka Tunggal Ika and inspire pride in the diversity that makes Indonesia truly special.

What it does

RagamNesia is an AI-driven language and culture app that brings Indonesia’s local languages, dialects, and folktales to life through truly immersive learning.

Speak authentically using our personalized text‑to‑speech feature, tailored for various tribes and dialects.

Learn through stories, with folktales transforming language lessons into interactive adventures.

Gain hands-on cultural insight, exploring traditions, expressions, and vocabulary shaped by regional heritage.

Track progress with ease, thanks to feedback loops and data-gathering tools aimed at refining user experience.

How we built it

We implemented a fine-tuned models and open-source speech synthesis systems like VITS to generate well defined Text-To_speech Model, that’s trained using Indonesian dialect .wav from open-SLR for Javanese, Sundanese, Balinese, Madurese, Makassarese to produces well defined model. For the language understanding and quiz generation or Language Identification we use OpenAI LLMs with prompt engineering to condition the LLMs to simulate tutoring and storytelling across different cultural settings.

We included folktales, expression and regional vocabulary that’s process and stored in MongDB cloud service that use loosely coupled document schema to provide easy, fast and reliable database. Thus providing Interactive quizzes, live TTS (Text To Speech) and thus give the opportunity to provide feedback loops that’s implemented using S3 bucket with it’s signed key.

Lots of time spent on building the AI model and trying to find the best to fit the native dialect speaker, undocummented API that’s rolled into chaos and lots of ideas poured into 1 application.

Challenges we ran into

Derick : I have never been involve inot business process with the government as out target Tristan Gautama : It's my first time integrating FE and BE, many mistakes happen William : first time implement AI text to speech technology... usually i'm a software engineer.

Accomplishments that we're proud of

We completed our backlog and planning.

What we learned

We learn many more things such as new tech, new way of discussing and solving problems mostly merge conflict 🙂

What's next for RagamNesia

Next... we would like to realize, perfected and deploy it :). Providing the easiest access of folklore, native speaking model and a chance of revitalizing our culture.