Expressive TTS
700 hours, full of feeling
Around 8 dedicated speakers record 700 hours — not flat studio reading, but speech carrying real emotion: laughter, lament, lullaby. Enough for a voice that actually sounds Kashmiri.
Flagship effort
Teach the moon
to speak Koshur
Zoon means moon. It is also the largest open speech corpus ever attempted for Kashmiri — built so the language can live inside the voice technology the rest of the world takes for granted.
The idea
A voice the machines have never been taught to hear — until now.
How Zoon is built · scroll
Two halves of one voice.
Speech-to-text
500 hours, 500 voices
500 hours gathered from about 500 everyday speakers — every district, dialect, age and accent — so recognition works for the whole valley, not just newsreaders.
Curate & align
Cleaned, labelled, aligned
Every clip is transcribed, timestamped and quality-checked — turning raw recordings into data a model can actually learn from.
Release openly
Free for everyone
The finished corpus is released openly for researchers, developers and dreamers — anyone who wants to build Kashmiri into their tools.
Your part
Add your voice
The STT half needs hundreds of ordinary speakers. Five minutes of reading aloud is a real contribution.
By the numbers
A corpus the size
of a language
0
hours of expressive TTS recordings
0
dedicated voices, full of feeling
0
hours of speech-to-text data
~0
speakers from across the valley
Lend your voice
Your accent belongs in the corpus.
No studio, no signup fuss — just your phone and your voice. Every dialect you add makes Kashmiri speech tech work for more people.