Paraspeak Decoding impaired speech in real-time

An AI-powered assistive device focused on impaired speech recognition. Designed to be radically affordable, simple, and accurate — so that more people can be heard.

Join the waitlist Partner with Paraspeak

As seen on

What are Speech & Voice disorders?

Disorders like dysarthria, presbyphonia and others make it difficult for people to form, pronounce or project words effectively. These conditions can result from stroke, paralysis, cerebral palsy, Parkinson’s disease, traumatic brain injury, or simply age-related changes. They create a significant communication barrier in everyday life that affects millions of people.

What is Paraspeak?

Paraspeak is a real-time speech recognition and enhancement system built to help people with a range of speech and voice disorders. It records atypical or impaired speech, runs it through a deep learning model, and plays out clear speech — all in a small, affordable, IoT-enabled device.

Paraspeak also focuses on under-researched languages and low-resource settings, making it accessible to those who need it most.

Person using Paraspeak device and smiling as their speech is understood

A congenital disorder patient who is happy that his voice has been understood correctly.

Awards & Recognition

2025 · Delhi, India

Samsung Solve for Tomorrow - National Winner

Winner in the Future of Health, Hygiene & Wellbeing category, awarded a seed fund and incubation at FITT, IIT Delhi.

2025 · Columbus, Ohio, USA

Regeneron ISEF – Fourth Grand Award

Representing Team India in the Biomedical Engineering Category at the world's largest pre-collegiate science fair.

2024–25 · Delhi, India

IRIS National Fair – Grand Award

Placed in the Top 20 nationally at India's largest research-based science fair, earning selection to ISEF.

Join the waitlist

Please note that Paraspeak is a research prototype. We're seeking early testers and clinical-trial partners, and are currently on the journey of scaling up this system into a mass-produced assistive device.

A chronic obstructive pulmonary disorder patient using the device.

Your Name (on behalf of the patient) Your Email (on behalf of the patient) Patient's Age Since when has the patient been suffering from this disease?

Native Language City Please briefly describe the diagnosis:

It’s helpful to describe the patient’s speech patterns. For example:

Speech can be formed but is slurred or unintelligible
Limited speech formation, a large amount of stuttering
No speech formation

If possible, specify if any aphasia or apraxia has been diagnosed.

By submitting this form, you acknowledge that the information provided will be used solely for the purpose of research, and seeking out early testers and clinical trial participants. No personally identifiable information about the patient is required as a part of this form, and any details you share will be kept strictly confidential. We do not sell, share, or disclose this information to third parties, and all data will be handled in accordance with applicable privacy and ethical guidelines.

The Founder

Pranet Khetan is an award-winning young innovator and researcher — a Grade 11 student from Shiv Nadar School Gurugram — who founded Paraspeak to revolutionize speech-assistive AI for people with dysarthria.

His work on Paraspeak has been recognised nationally and internationally, as a National Winner of Samsung Solve for Tomorrow 2025 (leading to Paraspeak’s incubation at FITT, IIT Delhi), and as a representative of Team India at the Regeneron International Science and Engineering Fair (ISEF) 2025 held in Columbus, Ohio, USA, where he won the 4th Grand Award in Biomedical Engineering. Pranet also presented his research at IIT Delhi’s EMPOWER Assistive Technology Research Conference as the only high-school researcher among professors, professionals, and college students.

Beyond Paraspeak, he has led teams in competitive robotics, advanced AI projects, and computational neuroscience research. A perfect SAT scorer who is also Trinity College London Speech & Drama certified, Pranet brings together technical excellence, leadership, and creativity to design impactful solutions at the intersection of AI, healthcare, and human communication.

Tech Experience:

Python MATLAB C/C++ EasyEDA ESP32 Fusion360 Web Development

Partner with Paraspeak

Paraspeak is more than a research prototype — it’s a mission to restore voices. We’re seeking partners from across healthcare, academia, and social impact to help scale this system. You can contribute in several ways:

Healthcare & Rehab Centres: Run pilot programs, support clinical validation, and connect us with patients in need.
NGOs & Community Groups: Collaborate on speech data collection, awareness campaigns, and outreach in low-resource settings.
Universities & Researchers: Contribute to trials, language research, and cross-disciplinary innovation.
Investors & Foundations: Support funding, scale-up manufacturing, and distribution for affordability and global reach.
Volunteers & Advocates: Help with local language datasets, patient engagement, or spreading the word.

Whether you’re a clinic, research group, nonprofit, or individual advocate — we’d love to explore how we can work together on the mission to restore the voices of those with voice and speech disorders.

Get in Touch

pranet@paraspeak.in