Talk instead of type.

One of the biggest game changers for me isn’t an AI scribe. It’s a speech-to-text dictation tool. I use it whether I’m writing a record or not, and I’d tell pretty much anyone who does anything on a computer to try one.

I use Wispr Flow. It works great. Superwhisper is also good. There are a lot of good ones. None of them pay me anything.

That’s the whole pitch. You talk. It types.

Why does speech to text beat typing for veterinarians?

Because people get their point across better speaking than typing. It’s faster, and you say things more fully and more naturally.

When I type an owner email, I shorten. I skip the sentence that would have made it clear. When I talk it, the whole thought comes out, the way I’d leave it on a voicemail. Same for a ChatGPT prompt. Same for an SOP. The extra context is the useful part, and talking is how you actually put it in.

How do I use dictation software in my clinic?

I put the cursor where I want the words. I hold down Control + Space. I talk. I let go.

It listens, cleans up the text, knocks out the ums, fixes the punctuation, and drops a near-perfect transcription into whatever app I’m in. Email. ChatGPT. Notes in Shepherd. A document. If I can type there, I can talk there.

The hotkey is just what I set mine to. Yours might be different. The loop is the same.

How does dictation software for vets actually work?

These tools all do roughly the same four things.

1. A push-to-talk hotkey starts the mic.
2. Speech recognition turns what you said into words.
3. An AI cleanup pass removes filler, fixes punctuation, and turns spoken mess into something you’d actually send.
4. The result pastes into the active text field.

That’s why it works in so many places. It isn’t plugged into Gmail, or ChatGPT, or your practice software. It’s typing where your cursor already is.

What do I actually use it for?

Owner emails. I’d rather talk the email than draft it. I still read it before I send it.

Prompting ChatGPT and Claude. The useful prompt is the real case, with the messy parts left in. That’s the whole point of the prompts article. Talking is how I get that much context in between rooms.

SOPs and staff notes. “Here’s how we do a dental discharge” said out loud is a better first draft than anything I’d sit down and write.

And regular notes, the ones that aren’t the exam. A reminder to myself. A message to the team. Anything I’d otherwise leave short because typing is a chore.

Is this a replacement for an AI scribe?

No. It complements one.

An AI scribe listens to the exam and writes the SOAP and the discharge. Dictation is for everything else: the email, the prompt, the SOP, the box you’re already in. I still run a scribe in the rooms. I still talk into Wispr Flow at the desk.

If you’re choosing a scribe, I wrote that guide. This is the tool I wish somebody had told me to install first.

What should I watch out for?

Read it before you send it. Near-perfect is not perfect. Names, doses, left and right still deserve a look.

And be mindful of the room. These tools hear whatever the mic hears. A conversation in the treatment area, an owner in the hall, a name you wouldn’t put in an email. Treat the hotkey like an open mic.

Look at the privacy settings of whichever one you pick. Wispr Flow says on its site that you choose whether your data is used to help improve the product. Superwhisper says it can work offline. I don’t know every vendor’s setup, and they change. Open the settings and read them.

I’m not selling either one. Use whatever works on your computer.

Where to go from here

Install one. Spend a day talking instead of typing. Then, if you want the exam-room version of this, that’s the scribe article. If you’re going to dump more context into ChatGPT, read what’s safe to paste first.