StarWhisper lets you write emails, documents, messages, and any text entirely by voice — without cloud upload, without subscriptions, without your data leaving your computer. Built for users with RSI, carpal tunnel, dyslexia, or any condition that makes typing difficult.
Accessibility voice typing exists because keyboards are not a neutral interface. For tens of millions of people — those living with repetitive strain injury, carpal tunnel syndrome, multiple sclerosis, Parkinson's disease, dyslexia, ADHD, or motor control conditions — conventional typing imposes a tax that able-bodied users never have to pay. Every document, every email, every form requires either physical pain, exhausting compensatory effort, or accepting a slower, degraded output.
Repetitive strain injuries affect approximately 1.8 million workers in the US annually according to the Bureau of Labor Statistics. RSI is not just discomfort — advanced cases cause chronic tendonitis, carpal tunnel syndrome, and in severe cases permanent nerve damage. Many knowledge workers hit a wall in their 30s or 40s where continued keyboard use is medically inadvisable, yet their entire livelihood depends on producing text.
For people with dyslexia, the challenge is different but equally disabling in professional contexts. The cognitive overhead of translating thoughts into correctly spelled, properly sequenced text is exhausting and inconsistent — it doesn't reflect the intelligence or capability of the person at all, but it shapes every written output they produce. Voice dictation removes the orthographic bottleneck entirely: speak naturally, think clearly, produce professional output.
ADHD and executive function differences create yet another category of need. Typing enforces a slow, sequential output mode that conflicts with how many ADHD thinkers actually work — in bursts, non-linearly, with ideas arriving faster than fingers can follow. Speaking is three to four times faster than typing for most people, which means voice dictation keeps pace with thought in a way that typing fundamentally cannot.
Windows Speech Recognition and the built-in voice typing (Win+H) have improved but still require extended training periods to reach acceptable accuracy, and they handle technical vocabulary poorly. Dragon Professional is expensive, has a steep setup curve, and is locked to specific microphone configurations. Cloud-based voice typing tools raise legitimate privacy concerns for people dictating medical records, personal correspondence, or legally sensitive documents.
StarWhisper was built on OpenAI Whisper — a speech recognition model that reaches human-level transcription accuracy on many audio types without requiring user-specific training. For accessibility users, this matters enormously: you don't spend weeks teaching the software your voice before it becomes usable.
Traditional voice recognition systems like Dragon require user enrollment sessions — you read passages, train pronunciations, build profiles. For people with fatigue conditions, motor difficulties, or chronic pain, this setup process itself can be prohibitive. StarWhisper requires no training. Install, launch, speak. The Whisper model was trained on 680,000 hours of diverse speech — your voice is already within its training distribution.
StarWhisper's floating widget sits above every window and types your speech directly into whatever application is focused — email clients, word processors, web forms, spreadsheets, coding IDEs, chat applications. You don't need to learn a different interface for each application. The accessibility voice typing experience is consistent across your entire Windows environment, which reduces cognitive load and the energy cost of switching contexts.
Many accessibility users dictate deeply personal content — medical history, legal correspondence, mental health journaling, disability accommodation paperwork. The idea that this audio is being uploaded to and processed on third-party servers is legitimately alarming. StarWhisper processes everything locally on your device. Nothing leaves your computer. This is a structural privacy guarantee, not a policy promise that can change with a terms-of-service update.
StarWhisper lets you choose between multiple Whisper model sizes. On a lower-end laptop, the "small" model delivers fast response with good accuracy. Pro users on capable hardware can use "large-v3" for maximum accuracy — relevant for users who speak with speech differences, unusual cadences, or significant accents that benefit from a more powerful model.
StarWhisper shows a real-time inline preview of transcribed text before it's inserted into the target application. This gives users a chance to catch errors before they appear in the document — important for people who may find correction more burdensome than prevention. The preview reduces the overall editing burden that remains after dictation.
Here's how a paralegal with RSI uses StarWhisper through a typical workday:
Launch StarWhisper floating widget. Dictate responses to 12 emails directly into Outlook — no typing at all. Legal terminology (indemnification, subrogation, tortfeasor) is handled accurately without custom vocabulary training. Total keyboard time: zero.
Open a Word document. Dictate a 1,200-word motion summary at conversational pace — roughly 15 minutes of speaking. Review the inline preview, make two small corrections. The document is substantially complete before any real editing begins.
Dictate research notes and case updates into the case management system's notes fields directly via the floating widget. StarWhisper works in web applications too — no copy-paste required.
Dictate a time-sheet summary and billing notes. Total keyboard use today: minimal, only for application shortcuts and navigation. Wrist pain: manageable, versus severe pain on typing-heavy days. Productivity: unchanged from pre-RSI baseline.
This workflow pattern — where voice handles composition and keyboard handles only navigation/selection — dramatically reduces the RSI load while preserving full professional output capacity.
Privacy is not an abstract concern for accessibility users — it is a live practical issue. Voice dictation captures not just work content but often the most intimate details of a person's life: medical conditions, legal situations, relationship dynamics, financial struggles. People using voice dictation out of necessity rather than preference have not chosen to broadcast this content to cloud infrastructure.
StarWhisper does not upload audio. It does not transmit transcripts. It does not train on your data. The entire inference pipeline runs on your Windows machine. The software is free of the surveillance capitalism model that makes most "free" cloud transcription products ethically questionable — those products are free because your audio data has value to their operators.
Under the Americans with Disabilities Act, employers are required to provide reasonable accommodations that enable employees with disabilities to perform essential job functions. Voice dictation software qualifies as an assistive technology accommodation. StarWhisper's offline operation means it can be approved by corporate IT security without the complications of a cloud data processing review — making accommodation requests faster and less contentious.
For GDPR-governed workplaces in Europe, the same argument applies even more strongly: the complete absence of any third-party data processor eliminates the GDPR compliance overhead that cloud-based accessibility tools require.
StarWhisper is designed to be immediately usable — accessibility software should not itself be an obstacle. Here's the recommended first-time setup:
Download from starwhisper.ai or the Microsoft Store. No account creation. No configuration forms. Run the installer and the app opens immediately.
In Settings, select "small" if you have an older machine or want faster response. Select "base" for even faster (slightly less accurate) performance on very low-end hardware. Pro users should use "large" for maximum accuracy — especially beneficial for users with atypical speech patterns.
Enable the floating widget mode in Settings. Position it in a screen corner that doesn't obscure your work. This is the core accessibility interface — click once, speak, release. Text appears in whatever application you're working in.
Set a global hotkey to start/stop dictation. A single key like F8 or a thumb-button on a mouse works well for users who want to minimize keyboard use. The hotkey works system-wide, so you can trigger dictation from any application.
Dictate a paragraph in your normal voice. Review the result. If accuracy is lower than expected, try the next model size up, or check microphone placement (closer is usually better for speech recognition). Most users reach satisfactory results on the first session with no additional configuration.
The ROI calculation for accessibility voice typing differs from standard productivity software. It's not just about minutes saved — it's about what becomes possible or impossible at all.
| Reduced daily output (pain limits typing) | 30–50% capacity |
| Physiotherapy / treatment costs | $100–300/month |
| StarWhisper Pro (unlimited dictation) | $10/month |
| Accessible even on free plan | $0 |
For many users, voice typing isn't a productivity feature — it's the difference between being able to work at all and not. The $10/month Pro cost is negligible against any reasonable estimate of the value of maintained productivity. And the free plan's 500 words/day covers light-to-moderate accessibility needs without any cost.
See also our guide on dictation software for writing and freelancer dictation software for more on professional use cases.
"I have severe carpal tunnel in both wrists after 15 years of software development. StarWhisper has let me keep working full-time. The floating widget works beautifully with Visual Studio Code — I dictate comments, docstrings, and prose sections while still using keyboard for code. The offline processing was non-negotiable for my employer's security policy."
— Senior Software Engineer, living with bilateral carpal tunnel syndrome
"As a dyslexic writer, typing was always a humiliating bottleneck — the gap between what I could think and what came out of my fingers was enormous. With StarWhisper, I dictate at the speed of thought. My writing has genuinely improved because I'm no longer self-censoring to avoid words I can't spell reliably."
— Freelance copywriter, dyslexia
"I was diagnosed with Parkinson's three years ago. My handwriting is gone and my typing has become unreliable. Voice dictation is now my primary way of communicating in text. StarWhisper handles my slightly slurred speech better than anything else I've tried — the larger Whisper models are genuinely impressive."
— Retired educator, Parkinson's disease
Better than most alternatives. The OpenAI Whisper model was trained on diverse audio including non-standard speech. The larger model sizes (medium and large, available on Pro) perform better on atypical speech. Some users with significant speech differences report needing to slow down slightly or repeat phrases, but the baseline accuracy without any training is higher than older ASR systems.
For text composition, yes — StarWhisper can produce text in virtually any application without any keyboard input needed for the text itself. Navigation (moving between fields, scrolling, selecting text) still requires some keyboard or mouse input. For users who need complete keyboard-free operation, StarWhisper pairs well with Windows' Eye Control or Switch Access features for navigation.
Yes. StarWhisper's free plan includes 500 words per day with no account required. For users with lighter dictation needs, the free tier may be sufficient indefinitely. The 500-word limit covers several emails, short documents, or a portion of a workday's writing needs.
StarWhisper operates independently of Windows Accessibility features and should not conflict with Narrator, Magnifier, or other assistive technologies. The floating widget works alongside any other accessibility tool running simultaneously.
Yes. StarWhisper supports 29+ languages including French, Spanish, German, Chinese, Japanese, Portuguese, Russian, Arabic, Korean, and many others — all offline. Set the transcription language in Settings for best results, or use auto-detect for multilingual speakers.
StarWhisper's complete offline operation makes it significantly easier to approve than cloud-based alternatives. There is no outbound data transmission for audio or transcripts. IT security reviews typically focus on network activity — StarWhisper's offline architecture means there's nothing to review in that respect. It's a locally-running application with no cloud dependencies.
Dragon requires a user enrollment process, is tied to specific hardware/microphone setups, costs hundreds of dollars, and has a complex interface. StarWhisper requires no training, no enrollment, and costs $10/month unlimited. Dragon has more advanced voice commands for navigating applications (select all, go to end of document, etc.) — StarWhisper currently focuses on transcription accuracy rather than full voice control. For users whose primary need is text composition, StarWhisper is faster to set up and more cost-effective. For full desktop voice control, Dragon remains the more comprehensive solution. See our professional transcription software comparison for more details.
StarWhisper is free to download and runs entirely on your Windows computer — no account required, no training needed, no audio uploaded anywhere. If you're living with RSI, dyslexia, ADHD, Parkinson's, or any condition that makes typing difficult, this is voice dictation software built to be genuinely useful from the first minute of use.
Windows 10/11 · No training · 100% offline · Free plan available · 29+ languages