For clinicians
Dysarthria technology for speech-language professionals
DysVoxa is a desktop assistive speech tool for Windows and Linux, designed for people with dysarthria and other speech impairments. It is not therapy or a medical device; it is a communication aid that may fit alongside clinical care and AAC planning.
Where DysVoxa fits in clinical practice
DysVoxa does not diagnose, treat, train speech, or replace SLP intervention. Instead, it addresses a practical gap: helping clients be understood in real-time conversations when their natural speech is difficult for listeners to parse.
Clients who may benefit
- Adults with dysarthria from stroke, Parkinson's, ALS, ataxia, cerebral palsy, or TBI
- People whose speech is affected by medical treatments or conditions
- Clients with adequate language and cognition who are frustrated by intelligibility barriers
- Users who need support in phone calls, video meetings, or group conversations
Clinical contexts
- As a compensatory tool alongside speech therapy
- For telehealth appointments where clear speech is critical
- To reduce communication breakdown during medical consultations
- As part of a broader AAC or assistive technology evaluation
How the DysVoxa workflow works
1. Client records a phrase
User-controlled start/stop - no continuous listening. Fit depends on the user's speech, environment, and microphone quality.
2. On-device transcription
Speech recognition runs entirely on the user's Windows or Linux computer. No audio leaves the device - important for clinical privacy and GDPR contexts.
3. Local text improvement
A local text improvement workflow is designed to reduce recognition errors typical of motor speech conditions.
4. Clear spoken output
Synthesized speech plays through speakers or a virtual microphone - usable in Zoom, phone calls, and telehealth platforms.
Current development stage
DysVoxa is a working desktop beta for Windows and Linux, tested with ataxic dysarthria. It is not yet a clinically validated medical device. Here is what is true today:
Available now (Free beta)
- Phrase-by-phrase recording with user-controlled timing
- On-device ASR and TTS - no cloud dependency for core functions
- Local text improvement shaped by dysarthric speech transcripts
- Voice profiles for personalized misrecognition patterns
- Virtual microphone output for telehealth and call platforms
- Guided setup - no command line or manual configuration
Planned (Pro tier)
- On-device ASR model tuning from user voice inputs
- Deeper personalization for returning users
- Expanded clinician and partner options
- Formal clinical evaluation and validation studies
Important clinical limits
- DysVoxa is an assistive communication aid, not a medical device or therapy program.
- Output quality depends on the user's speech, microphone, background noise, and setup.
- Users should review important messages before sending or speaking them when accuracy matters.
- DysVoxa should not be the only communication method for urgent or safety-critical situations.
Privacy and data handling
DysVoxa is designed with privacy as a core requirement, which matters especially in clinical and telehealth contexts:
- Speech recognition and text-to-speech run entirely on-device
- Biometric voice data and audio never leave the user's computer
- No telemetry or behavior tracking in the current beta
- Optional cloud text improvement sends recognized text only (never audio) and requires the user's own API key
- GDPR documentation and consent frameworks are part of the funding roadmap
See DysVoxa in action
Watch video walkthroughs of the DysVoxa workflow - from first launch to spoken output - to evaluate whether it may be relevant for your clients.
Watch the dysarthria speech demoQuestions or collaboration
If you're a speech-language pathologist, rehabilitation professional, or researcher interested in DysVoxa, we welcome your questions and feedback.