Dictation: Say It Instead of Typing It

You can say a message instead of typing it. Hold the mic control, talk, let go, and what you said arrives as text in the composer, where you can read it, fix it, and send it like anything else you typed. Dictation is input only. Donna reads it the same way as typed text and answers in text.

​

Where the control is

While you are recording, the control shows that it is listening. When you let go it shows that it is transcribing, then the text appears in the composer. Recording also stops on its own after about three seconds of silence, so trailing quiet is not recorded. If you tend to pause mid-thought, record in pieces: each piece is added to what is already in the composer, never written over it. The start list (⌘K) has the same mic next to its field, so you can dictate what you are about to work on there too.

​

What happens to what you said

The recording goes to Donna’s transcription service and comes back as text, lightly tidied: filler sounds are trimmed and your wording is kept. If the tidying step fails, you get the raw transcript instead of nothing. Transcription is also primed with the names of your active Outcomes and Actions, so the words specific to your work come back spelled the way you spelled them. Then:

1. The text lands in the composer. It is not sent yet.
2. You can edit it, add to it, or delete it.
3. You send it the way you send anything else.

Nothing is stored as audio. What persists is the text you send, which becomes part of the conversation like any typed message, with the same memory of your Outcomes and your patterns behind it.

​

When dictation helps

• The thought is long and tangled and typing it would flatten it into a to-do.
• You are avoiding something and talking through the first step is easier than writing it.
• You are wound up and saying it out loud costs less than composing it.
• Your hands are busy.

Speak plainly. There is no phrasing to learn and no command words to remember.

​

What dictation is not

• Donna does not answer out loud. There is no spoken reply, no read-aloud, and no audio version of anything Donna says.
• There is no live call. You cannot have a back-and-forth spoken conversation with Donna. A live voice mode existed earlier and was removed.
• There is no voice demo before you sign up. The browser demo that used to sit on the landing page, and the conversation it carried into a new account, are gone. Signing up starts you on the same first-run flow as everyone else.
• Dictation does not send for you. Every dictated message waits in the composer until you send it.

​

If it does not work

The browser asks for microphone permission the first time. If you decline, the mic control cannot record until you grant it in your browser’s site settings and reload. If the mic cannot start, or a transcription does not come back, a short note says so and the composer stays typable the whole time. Anything already in the composer is untouched; record again or type instead. If a recording comes back empty or wrong, edit the text in the composer or record again. Nothing has been sent yet, so a bad transcript costs you only the retake. Dictation has a daily limit. Past it, a note asks you to pick it back up later; typing is never limited.