If typing hurts, dictation is usually the first thing people try, and it's usually only part of the answer. Text entry is one problem. Navigation, clicking, selecting, and shortcuts are another, and most dictation apps don't touch them.
This guide covers both, in the order you should tackle them.
This is a practical software guide, not medical advice. See a clinician about the injury itself.
Before buying anything, turn on the tools already on your machine. They're free, they work offline, and they'll tell you within a day how much of your work voice can absorb.
macOS and iOS. Apple Dictation for text - free, on-device on Apple Silicon, with automatic punctuation since 2026. Apple Voice Control for everything else: navigation, clicking, corrections, full device control, free and on-device, with natural language input arriving via Apple Intelligence in 2026.
Windows. Win+H for voice typing - free, though cloud-based with no custom vocabulary. Voice Access on Windows 11 22H2+ for device control. Note that the older Windows Speech Recognition was deprecated in December 2023 and removed from current builds; Voice Access replaced it.
Android. Gboard Voice Typing for text, offline with a downloaded model. Voice Access for full device control, also offline with a language pack.
Linux. Numen for hands-free control, free and fully local. Vocalinux or Handy for dictation.
A day with these tells you whether the problem is text entry alone or the whole interaction model.
If dictation covers most of your writing, the next win is usually better text entry rather than more voice.
Good options once you've established the baseline: VoiceInk at $29 one-time on Apple Silicon, with per-app modes and custom vocabulary. superwhisper across macOS, Windows, and iOS, free on small local models or $249.99 lifetime. Handy free and offline on all three desktop platforms. Wispr Flow at $12 a user per month if cloud processing is acceptable and you want the least editing afterwards.
The feature that matters most for RSI is AI cleanup, because editing is typing. A tool that produces text you don't have to fix is worth more here than one that's marginally more accurate. Wispr Flow is the best at this; Amical does it locally and free by pairing Whisper with Ollama.
See the best dictation software for the full field.
If you need to stop using the keyboard entirely, dictation apps aren't enough. You need a command system.
Talon Voice is the strongest option available: free, local, macOS, Windows, and Linux (X11), scriptable in Python. Beyond speech it supports noise recognition - a pop or hiss as a click substitute, which saves your voice enormously - and eye tracking for pointing. Together, those three inputs let people work productively with no hands at all.
The community is the reason it works. Shared configurations, an active wiki, and people who have been doing this for a decade. Start from someone else's config rather than a blank file.
Budget two to four weeks to basic fluency. The alphabet comes first - a distinct word per letter, because speech models can't tell "b" from "d" from "p."
Numen does the same job on Linux, free and fully local.
Built-in alternatives - Apple Voice Control and Windows Voice Access - are less powerful but require no learning curve and no setup. For many people they're enough, and they're the right thing to try before committing weeks to Talon.
If your work is code, plain dictation won't do it. Reading code aloud produces garbage.
Talon Voice plus Cursorless is the most capable setup: Cursorless paints coloured markers on every token in VS Code so you can address any of them in a single command, and structural commands operate on functions and arguments rather than characters. Weeks of learning on top of Talon's own, and worth it if this is your job.
Serenade is much friendlier to start with, but its original project is unmaintained - check the community fork's activity before relying on it. Dragonfly and Caster are the older Windows-oriented path.
See the best voice coding tools.
This is the part that gets left out of most guides, and it's the most common reason people abandon voice as an accommodation.
Substituting voice for hands moves the load rather than removing it. Sustained daily dictation can produce vocal fatigue, hoarseness, and in some cases longer-term problems. People who switch abruptly from eight hours of typing to eight hours of talking often find themselves with a second injury.
Practical mitigations:
If you get hoarse, stop. A speech and language therapist is the right referral, and vocal technique is learnable.
Layer the inputs. Voice for text, eye tracking or a trackball for pointing, noise recognition or a foot pedal for clicking. Spreading the load across several channels is what makes this sustainable.
Prefer offline tools. If voice is your primary input, a service outage stops your work entirely. superwhisper, VoiceInk, Handy, Talon, and the built-in options all run locally. See the works offline collection.
Automate the repetitive. Snippets, text expansion, and per-app modes remove utterances entirely. Every phrase you don't have to say is voice you don't spend.
Have a fallback. Bad throat days happen. Keep a keyboard you can use briefly.
Free covers a surprising amount: Apple Dictation, Apple Voice Control, Windows Voice Access, Gboard, Voice Access, Talon Voice, Handy, Numen, Vocalinux.
Paid options worth the money: VoiceInk at $29 one-time, superwhisper at $249.99 lifetime, Wispr Flow at $12 a month.
Voicy offers 20% off for users with a disability, and 20% for students - worth asking other vendors whether they do the same, since several will discount quietly.
If the injury is work-related, occupational health or your employer's accessibility budget may cover hardware and software. Many organisations have a process for this that isn't advertised. In the UK, Access to Work funds assistive technology for employees; other jurisdictions have equivalents.
Start with the free built-in options - Apple Voice Control or Windows Voice Access - then add Talon Voice for full hands-free control, and a dedicated dictation app like VoiceInk or superwhisper for text.
Yes. Talon Voice with voice commands, noise recognition, and eye tracking supports fully hands-free work, and people do this professionally. Expect two to four weeks to basic fluency.
It can, particularly if you switch abruptly from typing all day to talking all day. Ramp up gradually, stay hydrated, speak quietly with a close microphone, and use non-speech inputs where possible.
Yes, and it's good. Apple Voice Control, Windows Voice Access, Android Voice Access, Talon Voice, and Numen on Linux are all free.
Yes. Talon Voice plus Cursorless is the most capable setup, and a number of developers work this way full time. See the best voice coding tools.
Often. Occupational health, accessibility budgets, and government schemes such as Access to Work in the UK commonly cover assistive technology. Ask - the process usually exists but isn't advertised.
Last verified: 10 August 2026.