Voice dictation

Most free-text fields in the wizard carry a microphone button for dictating instead of typing — your device’s own speech recognition, not an AI feature.

What it is

Dictation uses your device’s built-in speech-to-text: the native speech recognition engine on the installed app, or your browser’s own speech recognition when using the app on the web. It converts speech to text on the device (or through your platform’s own recognition service); it does not send anything to an AI model. The mic button appears beside single-line fields and pinned inside multi-line text boxes across most of the wizard’s observation, description and recommendation fields.

Before you start

On iOS, dictation needs the installed app build that includes the speech-recognition plugin and microphone permission granted — the web version of speech recognition doesn’t work reliably inside the app’s iOS webview, which is why a build without that plugin shows no working mic at all rather than a broken one.

How to do it

1. Tap the microphone icon next to or inside the field you want to dictate into.
2. Speak — the button turns red and the field shows “Listening…” while you talk.
3. Tap the button again to stop. Only your finished words are added to the field; nothing appears while you’re still speaking.
4. Continue typing or editing normally — dictation just adds text, it doesn’t take over the field.

How it integrates

Dictated text lands in the field exactly as if you’d typed it — it’s not tagged or treated differently anywhere downstream, including when it later goes through AI polish on fields that offer it.

Common problems

• The mic button isn’t there at all. Either this field doesn’t offer dictation, or your device has no working speech recognition available — on iOS specifically, check you’re on a build that includes the speech plugin.
• I get a permission error. Grant microphone access in your device’s settings for the app, then try again.
• Nothing was added after I spoke. Only a final, committed transcript gets added — if you stopped mid-sentence or the recognition produced nothing usable, try again and pause briefly before tapping stop.