A 15-year-old dream has come true today. I started a PhD with the dream of creating a system that chants any Sanskrit shloka perfectly.
And here I am opening sourcing - ṛ () ś-- -- (TTS) . This is the world's first vrutta-aware, open-source TTS for Sanskrit Chanting. I am making the model weights, training scripts, and even data (that I meticulously collected) public - https://prathosh.in/vagdhenu/
No large AI lab. No big engineering team. No venture-scale budget. Just a professor's conviction that one of humanity's oldest knowledge traditions deserves modern, open infrastructure.
The name comes from the Upanishadic phrase: "Vācaṃ dhenum upāsīta" - Like the mythical wish-fulfilling cow, Vāgdhenu is intended to make Sanskrit texts more accessible to students, teachers, researchers, and devotees everywhere.
Test out the live demo here and let me know your comments - https://prathosh.in/vagdhenu/ The entire system, from data collection to model building and demos, is built by a single person (your truly) using the powerful harness that we are building at LatentForce.
I have attached a sample audio file generated by the system.
P.S: Posting on behalf of my friend, their aren't on HN.
Comments URL: [https://news.ycombinator.com/item?id=48728732](https://news.ycombinator.com/item?id=48728732)
Points: 1