Your voice deserves
to be heard

Caunoi is AI-powered speech recognition for people whose speech is affected by stroke, ALS, cerebral palsy, Parkinson's and other conditions. Speak naturally — we'll make sure the world understands you.

Try It FreeRegister
200M+

people worldwide live with stroke, traumatic brain injury, cerebral palsy, Parkinson's, MS, or ALS and other motor neuron diseases — conditions that can make speech hard to understand.Sources

60M+

of them have speech difficulties, by our estimate — the people Caunoi is built for.How we estimate

$271B/year

in estimated earnings lost by family and friends caring for stroke survivors, who gave about 71 billion hours of unpaid care in 2021.Source

Lack of

personalized tools that help them both communicate and improve their speech

Who We Are

Caunoi is a product of KL Bridge. We build personalized AI speech recognition and speech practice for anyone whose condition makes their speech hard to understand, so they can be understood at home, with caregivers and in everyday conversations.

  • Stroke
  • ALS
  • Cerebral palsy
  • Parkinson's
  • Multiple sclerosis
  • Brain injury
  • And more
How It Works

Communicate today. Improve every day.

Caunoi combines real-time speech recognition with personalized practice — so your experience gets better the more you use it.

1

Set up your profile

Record a set of phrases so our AI learns your unique speech patterns. A caregiver can help with onboarding.

→
2

Talk freely

Speak into the app and get instant, accurate transcriptions — even when standard tools can't understand you.

→
3

Practice & improve

Daily guided exercises strengthen your speech while making the AI even more accurate over time.

Features

Built different from traditional tools

Unlike symbol boards and text-to-speech apps, Caunoi starts with your voice and makes it stronger.

🗣️

Personalized Speech Recognition

Our fine-tuned AI model adapts specifically to your speech patterns — not a one-size-fits-all approach.

🔊

AI Voice Enhancement

Amplify and clarify your speech output so others can hear and understand you in real conversations.

📈

Speech Practice & Therapy

Guided daily exercises that support speech therapy: rebuild speech after a stroke or brain injury, or keep it strong for longer with a progressive condition.

🔄

Gets Smarter Over Time

Every session feeds back into your personalized model, creating a virtuous cycle of improvement.

👨‍👩‍👦

Caregiver Friendly

Designed for family members and caregivers to assist with setup, track progress, and provide support.

🌏

Made for Everyone

Built for anyone whose speech standard tools struggle to understand, whoever and wherever they are.

About Us

Expert team. Real mission.

We combine AI research, business expertise, and clinical knowledge to solve a real communication gap.

Founder & CEO

Kien Nguyen, PhD

AI/ML researcher. PhD in Computer Science at University of Southern California.

LinkedIn →
Strategy Consultant

Loi Nguyen, MBA

MBA from University of California, Irvine. Former Autodesk and Nielsen. Expertise in go-to-market, sales strategy, and market research.

LinkedIn →
Medical Advisor

Tung Kim, MD

Specialist Doctor Level II in Rehabilitation at the National Geriatric Hospital. Lecturer at Hanoi Medical University.

LinkedIn →
🥈
Award

2nd Place — Vietnam AI Stars 2025

Caunoi was selected as a top finalist among 40+ teams in Vietnam's premier AI competition, recognized by leading experts from Google DeepMind, Yahoo, and the broader AI community for our work on personalized speech recognition.

Visit Competition Website →Read LinkedIn Post →
Partners

Supported by

  • NVIDIA Inception

    Member of NVIDIA's program for AI startups.

  • AWS Activate

    AWS cloud credits for training and running our models.

  • Lambda

    GPU cloud credits for model training.

  • Claude for Startups

    Member of Anthropic's startup program.

  • Speech Accessibility Project

    Approved access to the University of Illinois dataset of atypical speech recordings.

Ready to be heard?

Register to start using personalized AI speech recognition. We'll email you everything you need to get started.

Sources

200M+ people

200M+ is the sum of six published worldwide estimates of people living with each condition (about 208 million in total). The estimates come from different years, some people have more than one condition, and not everyone with these conditions has difficulty speaking.

  1. Stroke: 104.8 million (2023). Feigin et al., GBD 2023 global stroke fact sheet, Neuroepidemiology, 2026.
  2. Cerebral palsy: an estimated 50 million (2019) who could benefit from rehabilitation, modelled from GBD 2019 data. Cieza et al., The Lancet, 2020.
  3. Traumatic brain injury: 37.9 million (2021). Dong et al., GBD 2021 analysis, Medicine, 2026.
  4. Parkinson's disease: 11.77 million (2021). Luo et al., GBD 2021 analysis, Frontiers in Aging Neuroscience, 2025.
  5. Multiple sclerosis: 3.1 million (2026). MS International Federation, Atlas of MS.
  6. ALS and other motor neuron diseases: 0.27 million (2021). Peng et al., GBD 2021 analysis, Frontiers in Neurology, 2025.

60M+ with speech difficulties

Our rough estimate, not a published figure: each condition's number of people above, multiplied by the share with speech or language difficulties found in the studies below (about 60.5 million in total). For brain injury we count only moderate and severe cases (29.1 million). The studies cover specific groups and countries, so the true number is uncertain.

  1. Stroke, 25%: aphasia (23.7%) or dysarthria (27%) still present 3 months after stroke. Ali et al., International Journal of Stroke, 2015.
  2. Cerebral palsy, 36%: motor speech problems in 1,268 children. Parkes et al., Developmental Medicine & Child Neurology, 2010.
  3. Moderate/severe brain injury (29.1 million, 2021; Zheng et al., Medicine, 2026), 30% with dysarthria. Safaz et al., Brain Injury, 2008.
  4. Parkinson's, 51%: speech intelligibility clearly below that of healthy speakers. Miller et al., Journal of Neurology, Neurosurgery & Psychiatry, 2007.
  5. Multiple sclerosis, 44%: impaired speech or voice. Hartelius & Svensson, Folia Phoniatrica et Logopaedica, 1994.
  6. ALS, 80%: at some point unable to meet daily communication needs with natural speech. Beukelman et al., Neurology Research International, 2011.

$271B/year lost by stroke survivors' caregivers

Estimated earnings that unpaid caregivers (family and friends) of stroke survivors lost by working less or leaving work, and the hours of care they gave (1.37 billion a week, about 71 billion a year), worldwide in 2021, in 2023 US dollars. Across all 24 brain health conditions in the same study, caregivers gave an estimated 314.5 billion hours of care and lost US$1.7 trillion in earnings.

  1. Lastuka et al., The Lancet Public Health, 2026, table, with the publisher's correction of annual care hours.

Register for Caunoi

Tell us who you are and we'll email you how to get started.

Contact us

Questions, feedback or ideas? Send us a message and we'll get back to you.

Caunoi

Caunoi is a product of KL Bridge.
© 2026 KL Bridge. All rights reserved.