Before you start
You'll need all three of these:
An active CarePilot account
Dictate enabled for your organization and for your user account — these are two separate switches
An internet connection — required to sign in and to dictate
If either switch is off, Dictate tells you which one is missing so you know whether to contact your administrator or CarePilot Support.
Supported platforms
Platform | Hold this to talk |
macOS (Apple Silicon and Intel) | fn (globe) key |
Windows 10 and later | Ctrl + Space |
Dictation is English (US) in this release.
Install and set up
Download the approved build. Get it from GitHub releases or from your administrator, then install it like any other desktop app.
Sign in with CarePilot. Your system browser opens the CarePilot login page.
Follow the guided onboarding. It walks you through:
Permissions — microphone access, plus Accessibility on macOS
A microphone test — processed on your computer; no audio is saved
Your dictation style — Just talk, or Speak your punctuation (see below)
Your shortcut — the key you'll hold to talk
A short practice dictation
Personalization, including how long to keep your transcript history
There's nothing technical to configure. No models, engines, or audio settings to pick.
How to dictate
Click into the field where you want the text — your note, an email, a message box.
Hold the shortcut (fn on Mac, Ctrl + Space on Windows).
Speak at a normal pace.
Release the key. Your text appears where your cursor is.
Press Escape to cancel a dictation in progress.
The pill
A small pill sits at the bottom center of your screen and shows you exactly where things stand:
Ready → Listening (with live audio activity and a cancel button) → Transcribing → Inserting → Success or Error
The pill never steals your keyboard focus, so it won't interrupt what you're doing.
Hands-free mode
Hold-to-talk is the default. If you'd rather not hold a key down, you can switch to a hands-free toggle from the on-screen pill.
Always review before you sign. Dictate assists your documentation — you stay responsible for the final content. Read the inserted text before saving or signing.
Choosing your dictation style
Dictate is tuned for medical language in both styles. The only thing you're choosing is how you want to handle punctuation. Both produce identical finished text — they differ in what you say, not what you get.
Pick one during onboarding, and change it any time under Settings → Dictation.
Just talk (default, recommended)
Speak naturally and Dictate punctuates for you. Dates, doses, and phone numbers are formatted as you go.
You say: "patient presents with a headache nausea and vomiting"
You get: Patient presents with a headache, nausea, and vomiting.
One thing to know: in this mode, punctuation words are typed as words. Say "period" and your note reads period.
Speak your punctuation
Say your punctuation out loud and Dictate types it — the habit Dragon taught you still works.
You say: "patient presents with a headache comma nausea comma and vomiting period"
You get: Patient presents with a headache, nausea, and vomiting.
The spoken commands are:
period · comma · colon · question mark · exclamation mark · new line · new paragraph
Everything else is transcribed exactly as you said it.
Formatting options
These apply to both dictation styles:
Numbers as digits — "five hundred milligrams" becomes "500 milligrams"
Abbreviate units — "milligrams" becomes "mg", "milliliters" becomes "ml"
Filler words — "um" and "uh" are removed by default; turn this off to keep them in the transcript
Together: "um five hundred milligrams" → 500 mg
Personal vocabulary
Add your own terms — medications, provider and practice names, unusual spellings — and Dictate passes them to the speech service as recognition hints with every dictation, so they come back spelled correctly.
Find it under Settings → Personal vocabulary.
Your dictation history
Transcripts are stored in your CarePilot account, not on your computer. That means the same history follows you to any device you sign into.
Under Settings → History, choose how long transcripts are kept:
Don't keep them
1 hour
1 day
7 days
Indefinitely
With Don't keep them, the text disappears as soon as it's inserted into your app.
Your organization can set a maximum retention period. If it has, you'll see "Limited by your organization to…" and your choice is capped at that ceiling.
From the History view you can copy or delete individual transcripts. Delete all history removes every transcript in your account, across all devices — this can't be undone.
Audio is never kept, no matter what your transcript setting is.
Privacy and data handling
Speech is transcribed by a dedicated medical speech-recognition service over a secure connection. Audio is sent only for the duration of each dictation.
Dictate keeps no audio recordings. The onboarding microphone test is processed locally and nothing is saved.
Transcripts live in your CarePilot account and are deleted automatically after the retention period you choose.
Your sign-in credentials are stored in your operating system's credential store — Keychain on macOS, Windows Credential Manager on Windows — never in the app's web layer.
Access fails closed. If your access is revoked or expires, Dictate stops working.
What's in Settings
The main window (the pill is your primary interface day to day) covers:
Account and access
Dictation shortcut and pill behavior
Microphone
Dictation mode and formatting
Personal vocabulary
History and retention
Startup, updates, and diagnostics
Troubleshooting
Dictate records, but no text appears (macOS) Accessibility permission is missing. Go to System Settings → Privacy & Security → Accessibility and enable Dictate.
"Organization or account not enabled" One of the two access switches is off. Once your administrator makes the change, click Check again. If you're signed in with the wrong account, use Use another account.
"Can't reach CarePilot" in History Transcripts are stored in your CarePilot account, so History needs a connection. Check your internet and try again.
Audio sounds poor or playback degrades while the mic is active Bluetooth headsets are a common cause. Use your built-in microphone or a wired one instead.
Current limits
English (US) only in this release
macOS and Windows only — there's no official Linux build
An internet connection is required to dictate
For practice administrators
A few things that are yours to manage rather than the clinician's:
Enablement is two-level. Dictate must be turned on for the organization and for each individual user. A clinician seeing "Organization or account not enabled" is waiting on one of those.
You can cap retention. Set an organization maximum for how long transcripts are kept, and every user's personal choice is capped at your ceiling. They'll see "Limited by your organization to…" in their settings.
Distribute the approved build. Clinicians can install from GitHub releases, or you can hand them the build directly.
Permissions are per-machine. On macOS, clinicians need to grant Microphone and Accessibility permissions. Accessibility is the one most often missed — it's what lets Dictate type into other apps.
Audio is never retained, regardless of any retention setting.
Need help?
Contact your practice administrator for access and enablement questions, or reach out to CarePilot Support for anything else.
