Ask the assistant

In the Mac and Windows app, ask the assistant how to do something during a call, by typing or out loud. Only you see or hear it, and the room doesn’t hear you while you ask.

The Mac and Windows app has an assistant you can ask how to do something in Piano Gather while you're in a session: share your screen, fix your sound, change the layout. It answers from this help center and from how your own call is set up (whether you're the host, your microphone and camera, what you're sharing).

Only you see your question and its answer, and only you hear it read aloud. Nobody else in the call knows you asked.

Asking

Click Assistant at the top of the call (it's there on the setup screen too, before you join). Then:

Type your question and press Return, or click Ask.
Or ask out loud: hold the microphone button beside the question box, speak, and let go. You can also click it once, speak, and click it again. Your words appear as you speak. Cancel (or Escape) drops the question.

The keyboard shortcut is ⌘⇧K (Ctrl+Shift+K on Windows): press it to open the assistant and start listening, and press it again to ask what it heard.

Ask it to do something

The assistant can also do some things for you. Ask the way you'd ask a person: "turn my camera off", "put Mary in the spotlight", "mute everyone".

Your own settings happen straight away, and the assistant says what it did: your microphone and camera, switching camera, choosing a device by name ("use my USB mic", "switch my speakers to the AirPods"), Using headphones, Music mode, and opening the chat or the Asset queue.
So does running your own share. With the Asset queue open: "next video", "go back", "next page", "go to page 4", "zoom in", "fit the page", "make it quieter for everyone", "mute the video here". And "share my screen" (it opens the picker, where you choose what to share), "pause my share" and "stop sharing".
Anything that affects other people asks first. Click Do it (or press Return, or ⌘⇧K / Ctrl+Shift+K) to go ahead, or Cancel. If you're the host, that covers putting someone in the spotlight or taking it off, changing the layout, Mute all and Unlock mics, muting one person, starting or stopping a recording, inviting someone to share, removing someone from the session (you can still undo it from the bar at the top) or letting them back in, adding 5 minutes near the end, and making someone else the host. Anyone can ask it to start an alcove with someone ("start an alcove with Mary"; private unless you say open), to share the queue's item with the room, or to close and empty the Asset queue.
In an alcove, "back to the main call" and "we need more time" happen straight away.

Say a person's name as you'd say it in the room. If two people share a name, the assistant asks you to use the full name. It won't leave the session for you; use Leave.

Is my microphone muted while I talk to it?

Yes. While the assistant is listening, the room can't hear you. Your microphone is muted for everyone else from the moment you press the button until you let go, and the others see your mic-off badge. When you let go, your microphone comes back on. If you were already muted, you stay muted.

While an answer is read aloud, you stay muted too, until it finishes or you click Stop reading, so your microphone doesn't pick the answer up from your speakers and send it to the room. If Using headphones is on (under A/V), the answer is only in your ears, so you're not muted while it plays.

Note: Asking out loud isn't available while you're sharing a video or an app's sound (type your question instead), or while Hear yourself is running on the setup screen.

Answers read aloud

An answer to a spoken question is read aloud. Answers to typed questions are read aloud only if you turn on Read aloud at the top of the assistant. If you use a screen reader such as VoiceOver, it reads the answers instead.

Your privacy

On a Mac, a spoken question is turned into text on your Mac itself. Your voice isn't sent anywhere. The first time, macOS asks your permission to use speech recognition.
On Windows (and on the rare Mac that can't recognise your language by itself), a spoken question is sent to Microsoft to be turned into text as you speak. The assistant says so beside the microphone button.
Your question is answered by an AI service, together with the details of your call setup listed above. An answer read aloud is turned into speech by Microsoft. We don't keep your questions or the answers.

If the assistant isn't sure about something, it says so. You can always look it up here, or email matt@sonatina.com.