Research

Urdu Voice Cloner

AI voice cloning built specifically for the Urdu language โ€” starting from a poet's own recorded voice.

๐Ÿ”ฌ Research โ€” not a product yet
What it is

The language generic voice models get wrong

AI voice cloning built specifically for the Urdu language โ€” natural Urdu pronunciation, prosody and poetry recitation that generic voice models get wrong.

It was born from a project to preserve a poet's voice reciting his own verse. That is still what it is for, and it is why recitation โ€” not conversational speech โ€” is the hard case being solved first.

This is the earliest-stage work on this site. It is research, not a product with a shape yet.

AIUrdu-firstResearch
What's planned

What the research is aimed at

๐Ÿ—ฃ๏ธ

Natural Urdu pronunciation

Urdu pronounced as Urdu, rather than as an accent applied to a model trained on another language.

๐ŸŽต

Prosody

The rhythm and stress of spoken Urdu, which is where generic models most audibly fall down.

๐Ÿ“œ

Poetry recitation

Recitation of verse โ€” the case that prompted the project, and the hardest one.

๐Ÿ•ฏ๏ธ

Preserving a voice

Born from a project to preserve a poet's voice reciting his own verse.

Status

Where it stands today

Urdu Voice Cloner is not released. Nothing on this page is a promise of a feature, a date or a price.

ProgressEarly research

Want to hear when Urdu Voice Cloner is ready?

We invite a small circle of testers before every release. Write to us and say you're interested in Urdu Voice Cloner.

Get in Touch Back to the Roadmap