---
title: "Dictate instead of typing"
description: "Turn on dictation, speak into the message box with a click or a held key, and choose what happens to the finished transcript."
canonical: "https://bearly.ai/docs/features/dictation/"
markdown: "https://bearly.ai/docs/features/dictation.md"
---

# Dictate instead of typing

HTML: [https://bearly.ai/docs/features/dictation/](https://bearly.ai/docs/features/dictation/)

> Turn on dictation, speak into the message box with a click or a held key, and choose what happens to the finished transcript.

Dictation turns speech into text in the message box. You speak, Bearly transcribes, and the finished text lands in your draft so you can read it, edit it, and send it like anything else you typed.

Dictation is off until you turn it on. Nothing is recorded before that.

> **Note — Dictation and voice conversations are different**
> Dictation produces text for you to send. A [voice conversation](https://bearly.ai/docs/features/voice-conversations.md) is a spoken call where Bearly answers aloud. The dictation control is the sound-wave icon in the top-right corner of the message box; the voice control is the microphone icon below it.


## Turn on dictation



1. **Select the dictation button**

   It is the sound-wave icon in the top-right corner of the message box, labeled **Start dictation**. While dictation is off, selecting it opens **Settings** → **Chats** → **Dictation** instead of recording.

2. **Turn on Dictation**

   Use the switch beside the **Dictation** heading. The controls below it become available.

3. **Close settings and speak**

   Select the dictation button again. Allow microphone access when your browser or operating system asks. The button turns into a stop square, a status panel appears at the top of the window, and you hear a short rising tone.



Your dictation settings follow your Bearly account, so turning it on once covers your other devices.

## Start and stop recording

Select the dictation button to start and select it again to stop. You can also hold a key while you speak: keep <kbd>Right Shift</kbd> down, talk, and let go. A quick tap does nothing — hold the key for the recording to begin.

While you are recording, the status panel shows **Listening…** and the words as they are recognized. That text is a live preview; the version that reaches your draft is the finished transcript.

The keyboard shortcut works while the Bearly window is focused and needs a physical keyboard. On a phone or tablet, use the dictation button.

If you press any other key during a held recording, Bearly cancels it and inserts nothing. That is what keeps ordinary shifted typing from producing a transcript.

## Change the shortcut

Open **Settings** → **Chats** → **Dictation** and select the key shown beside **Keyboard shortcut**. Press the combination you want — up to six keys — and release. Bearly stores exactly what you pressed, so <kbd>Left Shift</kbd> and <kbd>Right Shift</kbd> stay distinct. The switch beside it turns the shortcut off if you would rather only use the button.

## Choose where the text goes

**After dictation** controls what happens to the finished transcript:

- **Fill the composer** (the default) adds the text to your draft. If the draft already has text, dictation is appended to it.
- **Send immediately** sends the transcript as a message as soon as it is ready.

If you finish a recording without saying anything, nothing is inserted.

## Clean up the transcript

**Transcript cleanup** is off by default. Turn it on and Bearly runs the finished transcript through a fast model that fixes punctuation and capitalization, and removes filler words and false starts. Its controls stay visible while it is off so you can see what it adds.

You can pick the **Model** and edit the **Instructions** that model follows. Cleanup runs on the text only — the original audio is not sent to it. Cleanup adds a moment before the text appears, and if it fails, Bearly uses the raw transcript rather than losing your words.

## Teach it your words

Under **Words and names**:

- **Vocabulary** is for names, acronyms, and technical terms you use — one per line or separated by commas. It guides cleanup toward the right spellings.
- **Replacements** are exact substitutions applied to every transcript, one `spoken → written` pair per line. `bear lee → Bearly` rewrites the spoken form wherever it appears as a whole word.

Replacements always apply. Vocabulary only affects the result when cleanup is on.

## Limits and availability

A recording stops on its own after 10 minutes and the transcript so far is delivered normally. For longer dictation, record in passes.

Dictation works in the Bearly web app, the desktop app, and the iOS and Android apps, wherever a microphone is available and permitted. On a managed team, an administrator can turn dictation off; the button then reads **Dictation is disabled by your team**, and the settings pane says your team has disabled audio transcription.

Live audio is sent through Bearly for transcription. Bearly does not store the recording.

## If dictation does not work

| What you see | What to do |
| --- | --- |
| Selecting the button opens settings instead of recording | Dictation is still off. Turn on the switch beside the **Dictation** heading. |
| **Dictation is disabled by your team** | Your team policy does not allow it. Ask a team administrator. |
| **Microphone access failed. Check your browser or system permission.** | Allow the microphone for Bearly in your browser's site permissions or your operating system's privacy settings, then try again. |
| **Transcription is unavailable. Try again.** | Bearly could not open the transcription connection. Check your network and start another recording. |
| **Dictation could not finish. Try again.** | The recording was interrupted and no text is inserted. Start a new recording; nothing from the interrupted one is kept. |
| The held key does nothing | Make sure the Bearly window is focused, that the shortcut switch is on, and that you are holding the key rather than tapping it. |
| Cleanup seems to do nothing | Confirm **Transcript cleanup** is on and that a **Model** is selected. Bearly falls back to the raw transcript when cleanup cannot run. |
| A replacement never applies | Check that the line has a `→` or `=>` between the spoken and written forms, and that the spoken form is a whole word in the transcript. |

## Continue through the documentation

- [Documentation index](https://bearly.ai/docs.md): Browse every Bearly guide.
- [Previous: Record and transcribe meetings](https://bearly.ai/docs/features/meetings.md)
- [Next: Talk with Bearly by voice](https://bearly.ai/docs/features/voice-conversations.md)