The shape of the day
You probably already know the rough arithmetic. Knowledge workers spend an average of 2.6 hours per day managing email alone, per a McKinsey Global Institute study (2012 figure - the proportion has only grown with Slack, Teams, and async messaging piled on top). That's 2.6 hours of keyboard time before you touch a document, a ticket, a pull request, or a code review.
For someone managing repetitive strain injury, that number isn't an abstract workday complaint. It's the daily tax on every wrist movement, every keystroke, every reach for the mouse. Research on IT professionals suggests 60% of people who spend 8+ hours a day at a computer will develop RSI symptoms at some point in their careers. That's not a fringe outcome - it's a near-occupational inevitability for developers, writers, analysts, and anyone else whose job is fundamentally a typing job.
RSI affects approximately 1.8 million US workers annually, with carpal tunnel syndrome alone comprising 32.5% of all occupational repetitive motion injuries and carrying the highest median days away from work of any occupational injury category at 25 days (per BLS data via Ergonomic Trends). And 57% of post-surgical CTS patients experience symptom recurrence within two years - which is to say, surgery often doesn't end the conversation. For a larger group - the ~39 million Americans with motor impairments from injury, neurological conditions, or congenital causes - typing isn't painful. It's just not fully available.
The goal of this article isn't to tell you that voice input is the answer to all of this. It's to walk through where voice input actually fits in a real workday, where it doesn't, and what AICHE does specifically for this persona.
Where typing slows you down
The keyboard burden for RSI and motor-impaired users isn't uniform. There are specific places in the day where it concentrates:
Long-form text output. Email replies, Slack threads, document drafts, PR descriptions, meeting follow-ups. These are sustained, continuous keyboarding. A 200-word email reply at 40 WPM (the average adult typing speed) takes about 5 minutes of active keyboard contact. At 150 WPM speaking pace - a figure confirmed by a Stanford/Baidu study that found speech roughly 3x faster than touchscreen typing - the same reply takes roughly 90 seconds.
Cross-device context switching. The person with RSI typically works on multiple surfaces: a desktop at a standing desk, a laptop in meetings, a phone for quick replies, maybe a tablet. Most voice tools are single-platform. That means re-configuring or rebuilding voice habits every time you move devices, or worse, falling back to typing on the devices the voice tool doesn't cover.
High-volume prompt writing for AI tools. Developers and knowledge workers sending prompts to Claude Code, Cursor, ChatGPT, or similar AI assistants generate a lot of text that is purely instrumental - it's not a creative output, it's an instruction. Typing those prompts adds friction with no value over speaking them.
Any text input that happens away from a keyboard. Walking to a meeting, standing in the kitchen, stepping outside. The RSI-affected body often benefits from breaks from a seated desk posture. Those moments are exactly when you want to capture a thought, reply to a message, or leave yourself a note - but also when you're furthest from your ergonomic keyboard setup.
How voice fits this workflow
Voice doesn't replace typing for everything, and it's worth naming where it doesn't work before talking about where it does.
Voice input is poor for:
- Precise cursor navigation (clicking, selecting a word three lines up)
- Short, highly structured data entry (filling in form fields, entering a date, changing a number in a spreadsheet)
- Anything that requires reading and responding simultaneously in a way that needs precise inline editing
For those things, your keyboard (or your OS voice-control navigation layer - more on that below) remains the right tool.
Where voice genuinely earns its place in a reduced-typing workflow:
Composing message bodies. You navigate to the text field with keyboard or mouse, place your cursor, and then speak the content. You're not removing keyboard use entirely - you're removing the highest-volume, most repetitive part of it.
Dictating document sections. Same model. Place cursor, speak a paragraph or a full section, let AI cleanup handle filler and punctuation.
AI agent prompts. If you're running Claude Code or Cursor, your prompts are often long and conversational. Speaking a 100-word prompt is faster and lower-strain than typing it.
Voice memos on the go. Capture a thought while stepping away from your desk. Let it transcribe and clean up in the background. Paste wherever it belongs when you return.
Watch + phone as capture surfaces. The Apple Watch in particular is relevant here - tapping the screen on your wrist to start a recording is a genuinely different ergonomic motion than sitting down at a keyboard.
AICHE specifically for RSI and accessibility users
Here's what actually matters for this workflow, with one-line explanations grounded in what the product does:
Global hotkey that works in any app, toggle-not-push. On Mac: ⌃+⌥+R. On Windows and Linux: Ctrl+Alt+R. Press once to start, press again to stop. You don't hold anything down. A single keypress begins the recording; another ends it. For RSI users, the distinction between "toggle" and "push-to-talk" is meaningful - sustained key holding is strain; two distinct taps are not.
Every platform, one subscription. macOS, Windows, Linux (.deb / .rpm / AppImage / Flatpak), iPhone, iPad, Apple Watch, Android, Chrome extension. Personal tier covers 3 devices; Pro covers 10. The multi-platform coverage matters because RSI users don't stay on one device - the voice habit needs to move with you, not be rebuilt per device. No mainstream voice-typing app in this category covers this full set.
Apple Watch recording. Tap your wrist to start a recording, tap to stop. No length cap. Syncs to iPhone and then to every other AICHE app. For someone who is stepping away from their desk to reduce strain, the Watch is one of the lightest-touch capture surfaces in this category.
Android home-screen widget. One tap on the home screen starts recording, another saves it. No need to open the app, navigate, press multiple buttons.
AI cleanup included on every recording. Filler words removed, punctuation added, paragraph breaks placed automatically. This matters specifically for voice-first users: raw dictation output requires editing. Editing by hand partially defeats the purpose if the goal is reduced keyboard time. Getting close-to-final text from a voice recording is the only outcome that actually lowers your total keyboard load.
Custom vocabulary (50 entries, synced across all platforms). Teach it your names, your product names, your technical jargon. This isn't about accuracy for its own sake - it's about reducing correction rounds after dictation. Every correction is a keyboard interaction. Fewer corrections is less keyboard contact.
Crash-proof save and offline queue. Record while away from your desk, on unreliable wifi, with the app backgrounded. The recording is saved locally and encrypted the moment you stop speaking. It processes when connectivity returns. The local queue reduces the risk of losing a recording when connectivity or processing fails. For someone who has shifted their workflow around not typing, that matters because reconstructing content often means falling back to the keyboard.
Voice Code for AI coding agents (Pro only). For developers with RSI: Voice Code is an opt-in continuous-listening mode - off by default, with a visible floating bar while active and a mute control you can trigger at any time. It lets you dictate prompts straight into Claude Code, Codex, Cursor, or Antigravity. Pause-aware auto-send means your prompt ships when you stop speaking - no Enter key, no clicking Send. Voice confirmations let you approve or reject agent actions by speaking. This is the piece that makes a nearly keyboard-free development workflow plausible - not because AICHE writes code, but because it handles the human-to-agent communication layer without requiring your hands. (Most AICHE users on desktop use the standard hotkey-toggle mode; Voice Code is a minority-adoption feature aimed specifically at AI coding agent workflows.)
REST API access (Pro only). If your workflow is custom - you want voice input to trigger a script, populate a specific tool, or feed into your own automation - the API gives you that without building your own transcription pipeline. This is the piece that lets accessibility-focused power users build exactly what their workflow needs.
Honest tradeoffs
AICHE is a voice-to-text layer. It's not a complete hands-free OS solution, and articles that pretend otherwise don't serve people who need accurate information.
AICHE does not control your operating system. It doesn't click links, scroll pages, open apps, navigate menus, or move your cursor. For a complete hands-free computer workflow, you need Apple Voice Control (macOS/iOS), Android Voice Access, or Talon Voice layered underneath. Those handle navigation. AICHE handles text input. The two are designed to work together, not compete.
Voice Code is a Pro feature. The continuous-listening mode for AI coding agents requires a Pro subscription ($9.99/mo or $8.33/mo on the annual plan). If you're on Personal and you want hands-free coding-agent prompts, you're using the standard hotkey-toggle dictation mode - which still works, just requires a manual keypress to start and stop each recording.
Audio processes via Groq, not locally. The transcription and cleanup pipeline involves a cloud round-trip. Audio is discarded immediately after processing, within 1 second. This is not a privacy concern for most users, but it is relevant if your use case requires air-gapped or strictly local processing. AICHE is honest about this; it doesn't claim local-only. For users who need fully local processing, there are tools like SuperWhisper (Mac only) - but local Whisper on a laptop is significantly slower (40-60 seconds for 15 minutes of audio vs. AICHE's ~3 seconds) and ships raw, unpolished output.
AICHE is not Dragon Professional. Dragon Professional Individual (~$699.99 one-time license) has spent decades tuning to enterprise dictation workflows, especially Windows-heavy environments. If you need deep Windows enterprise integration, a highly customized command vocabulary, or specialized profiles for long-form formal document dictation, Dragon is the established reference. AICHE is a better fit if you need multi-platform coverage, AI cleanup, lighter setup, and active development.
AICHE does not replace your phone's inline keyboard. The iPhone and Android apps are capture-first: you record a note, it transcribes, you copy the text. Apple Dictation (the microphone key on the iOS keyboard) handles inline dictation directly in text fields. AICHE and Apple Dictation are complementary, not the same tool.
Desktop app interface is English only. If you use AICHE on macOS, Windows, or Linux, the menus and UI are in English. Mobile (iPhone, iPad, Apple Watch, Android) supports 28 localized UI languages. Voice input and transcription work in 99 languages on every platform - the English-only restriction applies to the desktop interface, not to what you can speak.
AICHE is not a meeting transcription tool. If your goal is recording a full meeting and getting a timestamped, speaker-attributed transcript, that's a different category - Otter.ai is the established tool there. AICHE captures your own voice input and turns it into clean text. It does not join meetings, record other participants, or produce diarized transcripts.
No jurisdiction-specific guidance on RSI prevention or treatment. What reduces your strain, what exercises help, whether voice input is appropriate for your specific situation - that's between you and your clinician. This article is about workflow, not medical management.
What to try first
These are starting experiments, not a complete workflow rebuild. The goal is to find the substitution that saves the most keyboard contact for the least setup effort.
First experiment: Replace email body drafting with voice for three days.
Pick the 5-10 emails you write every day that are more than 3 sentences long. Navigate to the compose window, place your cursor in the body field, press the AICHE hotkey, and speak the email naturally. The cleanup pipeline will handle filler words and punctuation. You'll still type the subject line and recipient. After three days, count how many keystrokes you avoided. Most people find this is the highest-volume substitution available.
Second experiment: Use the Watch or Android widget for any thought capture away from your desk.
The next time you step away from your keyboard - to walk, stand, stretch - notice the thoughts that arrive and what you do with them. Instead of holding them in your head until you return to the keyboard or typing a quick note on your phone, tap your Watch or the home-screen widget and speak it. Let it process in the background. Come back to it when you sit down. The test is whether you actually miss fewer thoughts without adding a new strain source.
Third experiment (developers): Speak your next five AI agent prompts instead of typing them.
Open a Claude Code or Cursor session. Before typing your first prompt, activate AICHE with the hotkey, speak the prompt, stop the recording, let it insert. Do the same for the next four. Note whether the output quality differs from what you'd type and whether the reduction in keyboard contact feels material after 5-10 prompts. If you're on Pro, try Voice Code mode for the full pause-aware auto-send experience.
Try AICHE
7-day free trial, no credit card required. AICHE runs on the devices you already have - no new hardware, no retraining a model to your voice, no per-app setup. Install on the platforms most relevant to your workflow first, and expand from there.
View pricing and start free trial
Personal plan: from $3.99/mo (annual). Pro plan (includes Voice Code and API access): from $8.33/mo (annual).