Hold a key and talk
Dictation writes into the message box, so you read and fix the text before Bearly does anything with it.
Dictation is now in Bearly. It is the sound-wave button in the top-right corner of the message box: select it, speak, and what you said becomes the text of your draft.
It stays off until you turn it on. While it is off, selecting the button opens its settings rather than recording.
The instruction you would not have typed
You have just left a call about a slipped launch date. The spec is open in one tab, a half-written status note in another, and what Bearly needs is four sentences of context: who moved, what depends on it, and what the update should not say.
Typed, that is a minute of work, and you trim it as you go. Spoken, it is about forty seconds, and you say the whole thing — including the part about the customer who already saw the old date.
Then you read it. The transcript lands in the message box as an ordinary draft: cursor in it, editable, unsent. You fix the one name it heard wrong, delete a false start, and send.
That is the difference between dictation and a voice conversation. A voice call is something you listen to. Dictation gives you text to look at before Bearly acts on it.
Hold, talk, release
Selecting the button starts recording; selecting it again stops. There is also a hold shortcut: keep Right Shift down, talk, let go.
A quick tap does nothing. Pressing any other key while you hold cancels the recording and inserts nothing, which is what keeps a shifted capital letter from turning into a transcript. You can record a different combination of up to six keys — Bearly stores exactly what you pressed, so left and right modifiers stay distinct — or turn the shortcut off and use the button alone.
While you speak, a panel at the top of the window shows words as they are recognized. That preview tells you Bearly is hearing you. The finished transcript is what reaches your draft.
What happens to the text
By default the transcript fills the composer, appending to whatever draft is already there. If you would rather skip the review, After dictation can send it immediately.
Transcript cleanup is off by default. Turn it on and a fast model fixes punctuation and capitalization and removes filler words and false starts, working from the text alone — the audio is not sent to it. Its controls stay visible while the setting is off, so you can see what enabling it adds before you accept another step between speaking and seeing your words. If cleanup fails, you get the raw transcript instead of nothing.
Two lists handle the words transcription gets wrong for everyone in your line of work. Vocabulary holds names, acronyms, and technical terms. Replacements are exact substitutions, one spoken → written pair per line, applied to every transcript whether or not cleanup is on.
Dictation settings follow your account, so the shortcut and the word lists are the same on your other devices.
Where it runs
Dictation works in the Bearly web app, the desktop app, and the iOS and Android apps, wherever a microphone is available and permitted. The keyboard shortcut needs a physical keyboard; on a phone or tablet, use the button.
A recording stops on its own after 10 minutes, so long stretches need more than one pass. The shortcut works while the Bearly window is focused: it is not a system-wide hotkey and will not type into another application.
Live audio is sent through Bearly for transcription. Bearly does not store the recording. On a managed team, an administrator can turn dictation off.
The dictation guide covers the settings in full.