Voice Dictation
- Desktop, mobile and web
- All plans
What it does
Dictate your prompts instead of typing them. Click the microphone in the composer, speak, click again: the transcript is inserted in your draft at the cursor, ready to reread and send with Enter. Nothing is sent automatically. On desktop the audio is transcribed by the AgentsRoom voice backend (an accurate model for short clips, Whisper for recordings over 90 seconds, chosen automatically); on the phone recognition runs on-device and only the text is relayed, end-to-end encrypted, to the focused desktop agent. A live waveform shows the mic is picking you up, and silence is never transcribed, so a muted mic cannot produce invented text.
Where to find it
- Desktop: the Dictate button (microphone) in the composer toolbar of every agent. Right-click it for Global dictation settings and Project dictation settings.
- Desktop keyboard: Cmd+Shift+D (macOS) / Ctrl+Shift+D (Windows, Linux) starts dictating into the focused agent, Escape stops. Both are listed and rebindable in Settings > Keyboard shortcuts > Dictation, and shown in the mic tooltip.
- Desktop and mobile: the microphone next to the answer field of an agent's question card ("A question for you"). The transcript is added to the answer, nothing is sent until you press Enter or Send.
- Desktop and mobile: the same Dictate microphone next to the answer field of a decision on a backlog ticket, and in Decisions to take (since 2026-10-29). On the desktop it sits left of Withdraw and Decide, on the phone right of the field. The words go into Your answer when the decision has no options; on the desktop a decision with options dictates into Why (optional) instead. Nothing is recorded until you click Decide.
- Settings > Composer: dictation language, sounds, music pause, Offline dictation.
- Mobile companion: the microphone button on the agent screen (press and hold).
How to use it
- Put the cursor where the text should go and click Dictate. The first time, the OS asks for microphone access.
- Speak. The button pulses red with a live waveform. Click again to stop.
- The transcript is inserted at the cursor, split into one line per sentence. Edit if needed, press Enter to send.
Keyboard: press Cmd/Ctrl+Shift+D once to start, speak, then press Escape (the stop shortcut) to stop; the text stays in the draft, nothing is sent. Push-to-talk: hold Cmd/Ctrl+Shift+D while you speak and release it to stop. The start shortcut never stops a running take and the stop shortcut never starts one. It works in the main window, project windows and detached terminal or Assistant windows. With no agent open, a notice says so.
Per-project tuning: Project dictation settings lets you edit the context sent with each dictation (agent and project names are filled in; add domain terms, library or product names) and force a spoken language for this project.
Mobile: hold the mic, speak, release. The recognized text is pasted into the focused desktop agent's prompt, without Enter; you confirm from the phone or the desktop.
Settings
sttLanguage(global, Settings > Composer > Dictation language; project override in Project dictation settings):autoby default, or one of 21 spoken languages (Czech and Slovak added on 2026-10-12). Force one only when short clips get mis-detected.sttPromptContext(project, Project dictation settings): the context template sent with each dictation.voiceInputDevice(global, machine-local; Voice Mode panel gear > Microphone, or Settings > Composer > Microphone and speaker): the microphone used by Voice Mode and voice dictation.null= System default. A test button shows the live input level. If the device is unplugged, the system default is used until it comes back.dictationSounds(global): start and stop cues.stopMusicWhenRecording(global): mute the system audio output while recording, restored when you stop. On Windows it used to change the speaker channel volume instead of muting; fixed on 2026-09-24.localSttEnabled,localSttMode(fallbackoralways),localSttModel(tinyorbase) (global, Settings > Composer > Offline dictation): on-device Whisper, downloaded once (about 42 or 78 MB), used when credits or network run out, or always.composerHiddenActions(global): hide the mic button.dictationStartShortcut,dictationStopShortcut(global, account-synced, Settings > Keyboard shortcuts > Dictation): portable chords (Mod= Cmd on macOS, Ctrl elsewhere). Empty = default (Mod+Shift+DandEscape). The start shortcut needs Cmd, Ctrl or Alt (or a function key); the stop shortcut can be any key (Escape, Space…) because it only listens while a dictation runs. Use Reset to go back to Escape (pressing Escape in the recorder cancels the recording).
Agent tools (MCP)
None. Dictation writes into your draft; agents only receive what you send.
Providers
All providers, no difference: the transcript is text in the composer. The transcription model is chosen by AgentsRoom, not by the agent's CLI.
Mobile
The keyboard shortcuts are desktop-only (the phone has no keyboard; its mic is already press-and-hold). Present, with a different engine: the phone's own speech recognition (audio never leaves the device), text relayed encrypted to the focused desktop agent as a paste without Enter. If no desktop terminal is focused, the phone says so and sends nothing. The out-of-credits case falls back to the device engine.
Limits
- Requires a signed-in account for server transcription. With offline dictation set to always (
localSttMode=always), no account is needed: signed out, the mic opens the sign-in window only when a transcription has to go through the server. Included time per calendar month: 5 minutes of transcribed audio on Free, 30 minutes on Plus, 3 hours on Pro. These three values are compiled into the product: no back-office can widen them. Beyond the allowance each transcription spends voice credits (shared with Voice Mode and Read Aloud). Empty balance: dictation pauses and offers Keep dictating without credits (offline model) or a top-up. - Your own OpenAI key removes the credit spend, see Bring Your Own OpenAI Key.
- macOS: without microphone permission the recording is silent; the composer shows "Microphone blocked or no audio captured".
- Windows: when Let desktop apps access your microphone is off in the Windows privacy settings, the recording is refused up front and the same "Microphone blocked" notice appears (since 2026-09-23; before, the recording ran on silence and was dropped without a word).
- Recordings over 90 seconds use Whisper, which is more robust on long, noisy audio; repeated sentences are collapsed as a safety net.
- The offline model is less accurate than the online one.
Common questions
- Can I dictate with another microphone than the system default? Yes: Settings > Composer > Microphone and speaker > Microphone. The choice is per machine; Test the microphone shows the live level, and an unplugged device falls back to the system default.
- Can I dictate in Czech or Slovak? Yes, since 2026-10-12: pick Čeština or Slovenčina as the dictation language, or keep auto-detect.
- Offline dictation is on "always" but the mic asks me to sign in. Fixed on 2026-10-12: in that mode the mic works signed out. Update the app.
- The transcription failed: do I have to dictate again? No, since 2026-10-13 the recording is kept and the mic button turns into a retry button (amber, circular arrow). Click it to send the same audio again; Shift+click discards it and starts a new recording. On the phone: tap to resend, long press to discard. The kept recording is lost if you close the app.
- Can I dictate my answer to a decision? Yes, since 2026-10-29, on the desktop and on the phone: the Dictate microphone sits next to the answer field of the decision, on the ticket and in Decisions to take. The folded pill of a decision that has options shows no microphone, because it has no text field; open the decision to get it. See Decisions.
- Does dictation send my prompt? No, it only fills the draft. You press Enter.
- Is there a keyboard shortcut for dictation? Yes, since 2026-10-26: Cmd/Ctrl+Shift+D starts, Escape stops, and holding Cmd/Ctrl+Shift+D works as push-to-talk. Change them in Settings > Keyboard shortcuts > Dictation.
- Escape no longer reaches my agent while I dictate. Expected: during a take, Escape is the stop shortcut. Once the dictation stops, Escape works as usual. Bind another stop key in Settings if you need Escape in the terminal while dictating.
- Can I mix typing and dictation? Yes, the transcript is inserted at the cursor, not in place of the draft.
- Why does forcing a language give garbage? The model transcribes in the forced language whatever you speak. Keep Auto-detect unless detection fails on short clips.
- The mic looks broken and nothing is transcribed. Most often the voice credit balance is empty or the OS blocked the microphone. Hover the mic button: its tooltip shows the reason ("Microphone blocked or no audio captured", "No speech detected, nothing was transcribed.") instead of the usual title.
- The counter ran but no waveform and no text appeared (Windows). Fixed on 2026-09-23: when the audio analysis never started, the recording is now transcribed instead of being discarded as silence.
- Which model is used? Chosen automatically by duration; there is no model picker any more.
Related
- Voice Mode: two-way spoken conversation, same credits.
- Read Aloud: the output side, text read out loud.
- Bring Your Own OpenAI Key: keep dictating past the included time on your own key.
- Decisions: the decision answers you can dictate too.
- Scratchpad: the composer the transcript lands in.