10M+
People worldwide live with dysarthria
One of the most common and least-served motor speech conditions globally.
Free forever · Windows & Linux
Free voice correction & speech clarity for people with dysarthria.
DysVoxa records one phrase at a time, converts it to text locally, corrects recognition errors, and speaks a clearer, more understandable output - through speakers or apps like Zoom. Free forever, built by a founder living with ataxic dysarthria.
Voice data never leaves your deviceFree forever - no subscriptionWorks fully offline
- 100%
- On-device
- 0
- Voice uploads
- Free
- Forever
- Win + Linux
- Desktop
DysVoxa - Free foreverListening · phrase 1 of 1
Raw recognition
“cn you hear me on th call”
Corrected
“Can you hear me on the call?”
Speaking clear output…
1. Record
You control each phrase
2. Understand
Local transcription
3. Correct
Dysarthric-aware
4. Speak
Clear TTS output
Helping with speech affected by
StrokeParkinson'sALSCerebral palsyAtaxiaTraumatic brain injuryPost-surgeryMedication effectsMultiple sclerosisHuntington’sMyasthenia gravisBell’s palsyStrokeParkinson'sALSCerebral palsyAtaxiaTraumatic brain injuryPost-surgeryMedication effectsMultiple sclerosisHuntington’sMyasthenia gravisBell’s palsy
The problem
Thoughts are clear. Speech muscles aren't.
Tens of millions of people live with dysarthria - a motor speech condition where words come out slurred, slow, or hard to understand. The real cost goes beyond being misunderstood: after enough failed conversations, many people stop speaking up at all.
~1 in 3
Avoid calls & conversations after repeated breakdowns
The cost goes beyond being misunderstood - people withdraw from daily life.
$50/mo
Typical price of cloud-based alternatives
Existing tools require constant internet, send voice to remote servers, and lock out users who cannot pay. DysVoxa is free forever.
Few
Local-first, offline options exist today
Most solutions are cloud-only, Android-only, or closed to new users.
DysVoxa isn't therapy and isn't a cure. But if voice correction helps someone stay in one more phone call, one more appointment, one more conversation they would otherwise avoid - that's the point.
How it works
Four steps to clearer speech. All on your device.
DysVoxa runs a simple loop: record, understand, correct, speak. Every step happens locally on your Windows or Linux PC - no internet required after install.
- STEP 01
Record
You choose when each phrase starts and stops. No always-on listening - you stay in full control of every recording.
- STEP 02
Understand
DysVoxa transcribes your speech to text locally on your PC. No audio is ever sent to any server - recognition happens on your own hardware.
- STEP 03
Correct
Applies voice correction locally - fixing recognition errors common in dysarthric and unclear speech, guided by your voice profile if you’ve created one.
- STEP 04
Speak
Reads the result aloud with natural text-to-speech, routed to speakers, a virtual microphone for calls, or both - your choice.
Why DysVoxa
On-device voice correction for privacy & reliability
Most assistive speech tools are cloud-only, expensive, or platform-locked. DysVoxa takes a different approach: local-first processing, designed around dysarthric speech, and free forever.
Built for impaired speech
Not repurposed dictation software. DysVoxa’s correction is designed from day one for unclear, garbled, or hard-to-understand speech that standard tools fail to recognize.
Privacy-first by design
Speech recognition and voice output run entirely on your computer. Biometric voice data and audio never leave your device. No telemetry, no behavior tracking in the beta.
Works fully offline
Core features run locally, so the app works without internet after install - reliable exactly when communication matters most.
Personal voice profiles
Create a voice profile and DysVoxa can guide correction around the ways your speech tends to be misrecognized, handling repeated mistakes more intelligently.
Guided single-app setup
One desktop app with everything built in. Lists what’s installed, what’s missing, and downloads required components. No command line, no servers to configure.
Uses your natural speech
Not a symbol board or pre-programmed phrases. You speak naturally and DysVoxa improves your voice output to make it clearer and more understandable.
Built for real calls
A virtual microphone routes corrected output to Zoom, Teams, and other communication apps - so clearer speech reaches the people you actually talk to.
No voice cloning
DysVoxa uses synthesized text-to-speech for output. It never speaks in your recorded voice - keeping biometric identity safe by design.
How it compares
How DysVoxa compares to other assistive speech tools
Several tools exist for people with dysarthria and non-standard speech. Each takes a different approach to the same core problem: helping impaired speech be understood.
| Feature | Our appDysVoxa | AlternativeVoiceitt | AlternativeGoogle Project Relate | AlternativeApple Personal Voice |
|---|---|---|---|---|
| Approach | Speech-to-speech: records phrase, corrects locally, outputs clear TTS | Cloud AI trained on atypical speech patterns | Personalized ASR from 500+ recorded phrases | Voice banking + typing via Live Speech |
| Privacy | Fully local - audio never leaves device | Cloud-based - recordings stored on servers | Recordings uploaded to Google for training | On-device voice creation |
| Internet required | No - works fully offline after install | Yes - constant connection needed | Partially - training & Repeat need internet | No - works offline once created |
| Platform | Windows & Linux desktop | Web app (any browser) | Android only | Apple devices only |
| Cost | Free forever; Pro planned to cover support & engineering | $49.99/mo or $499.99/yr | Free (research, not accepting users) | Free (built into Apple OS) |
| Designed for | Dysarthria & motor speech specifically | Non-standard speech broadly | Non-standard speech broadly | Voice loss risk; not dysarthria |
| Use in calls | Yes - virtual mic for Zoom, Teams | Limited - in-person & dictation | No direct call integration | Yes - FaceTime & Phone via Live Speech |
| Availability | Open beta - email for access | Subscription + partner network | Closed - not accepting new users | On supported Apple hardware |
Voiceitt
VisitThe closest alternative
Voiceitt is the most established product in this space, with clinical trials and a global assistive-tech distributor network. It uses cloud AI trained on a proprietary atypical-speech database. The key differences: it requires constant internet, processes recordings on remote servers, and costs $49.99/month. DysVoxa is local-first and free forever - meaningful for users who prioritize privacy, offline reliability, or simply cannot afford a subscription.
Google Project Relate
VisitPromising but closed
A Google Research initiative that builds a personalized speech model from 500+ recorded phrases, offering Listen, Repeat, and a typing keyboard. It’s Android-only, requires uploading recordings to Google for training, and is currently not accepting new users.
Apple Personal Voice + Live Speech
VisitVoice banking, not correction
Creates a synthetic voice that sounds like you, combined with Live Speech for typing-to-speak. Privacy-friendly and free, but designed for people at risk of losing their voice entirely - not for correcting ongoing dysarthric speech. Requires Apple hardware and manual typing rather than speaking naturally.
Traction & roadmap
A working beta today. A clear path forward.
DysVoxa is a working desktop app for Windows and Linux in beta. The core loop - record, understand, correct, speak - has been tested with ataxic dysarthria. Here's exactly what works now and what's planned.
Shipping now
Works today (Free forever)
- Phrase-by-phrase recording - you control when each phrase starts and stops
- On-device speech recognition - no audio sent to any server
- Local speech correction designed around dysarthric speech patterns
- Voice profiles that guide correction around personal misrecognition patterns
- Natural text-to-speech output to speakers or virtual microphone
- Guided first-run setup - no command line, no manual file copying
- Quick phrases for frequent responses
- Optional cloud text improvement (text only, never audio) with your own API key
On the roadmap
Planned for Pro
- More personalized speech support
- On-device ASR model tuning using your voice inputs
- Improved personalization for returning users
- Expanded clinician and partner options
- Personalized voice-related features (future research)
- Adapter management for Pro users
Free forever is a commitment, not a trial. Pro is the planned paid tier - its revenue covers support and engineering costs only. The core app stays free for everyone, forever.
Free forever
Free now. Free forever.
DysVoxa Free is - and will remain - free for everyone who needs it. No subscription, no trial, no expiring beta. DysVoxa Pro is the only planned paid tier, and its revenue simply covers support and engineering costs so the project stays alive.
The core app will never be paywalled. Ever.
DysVoxa Free
Free forever€0/ always
The full beta - shipped today, free for life.
Request beta access- Guided first-run setup
- Phrase-by-phrase recording
- Local speech recognition
- Local speech correction built for dysarthric speech
- Voice profile to guide personal correction
- Output routing to speakers or virtual microphone
- Quick phrases for frequent responses
- Free forever - no subscription, no expiry, no paywall
DysVoxa Pro
PlannedTBD/ subscription
Deeper personalization. Paid only to cover support and engineering costs - not to paywall the core app.
Back development- Everything in Free, always included
- More personalized speech support
- On-device ASR model tuning with your voice inputs
- Improved tuning for returning users
- Expanded clinician and partner options
- Pricing covers support and engineering costs only
Free is a commitment, not a trial. DysVoxa Pro exists only to sustain ongoing support and engineering - the core speech correction app stays free for everyone, forever.
Invest & back us
Help bring accessible communication to the people who need it.
Funding covers the core build, a 12-month operating window, platform fees, and a contingency buffer. Every donation funds the work - and the project keeps the same privacy-first rule: your voice data never leaves your device.
Campaign progress
Core goal · stretch ceiling
Raised so far
€0
Goal
€28,575
ceiling €33,719
€0Goal €28,575Ceiling €33,719
Now: donations keep the beta moving
Hardware, testing, operating costs, and day-to-day work toward a public release.
After the goal: stretch funding
The stretch range is an extra €5,000-€5,500 for added polish and expansion.
Use of funds
How every euro is allocated across the campaign.
- Core build38%
Engineering the local speech-correction pipeline and TTS routing.
- 12-month operating window32%
Keeps the project running - hosting, tooling, and ongoing iteration.
- Platform fees12%
Campaign and payment processing fees on the funding platform.
- Contingency buffer18%
A reserve for hardware, testing, and unforeseen costs.
Privacy is non-negotiable. Funding accelerates development - it never buys the right to move voice processing off-device.
Built by someone who uses it
DysVoxa was built by a founder living with the condition it serves.
After diagnosis, everyday communication became an obstacle: repeating myself on calls, losing momentum in conversations, watching routine interactions become exhausting. With significantly limited mobility, reliable remote communication became essential - not optional.
The approach was pragmatic: reduce communication breakdowns, measure what works, and iterate until the system proves useful in daily life. DysVoxa is the result - designed for real communication needs, with local-first processing for privacy and reliability.
- Igor, Founder, SIA DysVoxa
Why it matters
Why speech improvement changes lives.
For many people with dysarthria or other speech impairments, the problem isn't just having hard-to-understand speech - it's the habit of withdrawing after too many failed exchanges. People stop speaking up: avoiding calls, appointments, and conversations they would otherwise have.
DysVoxa is not therapy and not a cure. But if voice correction helps someone stay in one more phone call, one more appointment, one more conversation they would otherwise avoid - that's the point.
Tens of millions affected
Dysarthria is one of the most common and least-served speech conditions worldwide.
Working beta, free forever
Your support turns it into a reliable public release - the core app stays free for everyone.
No server ever sees your voice
Speech recognition and text-to-speech run locally on your device, by design.
FAQ
Questions, answered.
Everything you might want to know about DysVoxa - what it is, how it works, and how to get access.
DysVoxa is a desktop voice correction and speech improvement app for people with dysarthria and others who need reliable, understandable voice output in daily life, therapy, and remote calls. It records speech one phrase at a time, converts it to text, corrects recognition errors, and speaks the result clearly with synthesized text-to-speech.
Get access
Try the beta. Join the mission.
DysVoxa Free is a desktop app for Windows and Linux, currently in beta - and free forever. Request access and we'll get you set up - your voice data never leaves your computer.
Minimum hardware
CPU 4 cores / 8 threads · 16 GB RAM
Below these specs DysVoxa may still work, but with higher latencies and slower processing.