Joins and speaker requests, announced aloud
Listeners who only listen are now picked up by the join announcement, and speaker requests are read out. Both are for noticing what participants are doing while you talk.
The update on 14 September 2026 made the join announcement pick up listeners who only listen, and added reading out speaker requests. Both exist so a host can notice what participants are doing while still talking.
Announcing listeners who never speak
The previous join notification could not always name someone who was only listening. Now people newly found in the Space's participant information are checked as well, and announced as "so-and-so has joined".
Because the detection runs through participant checks, the announcement does not necessarily play the instant someone arrives. People already known are not announced again. The join audio is not a roll call of everyone present.
When you hear one, try answering at the next break — a quick welcome, or a line about what is being discussed. It works best as a prompt to greet somebody who joined partway through. When there is no announcement, check the participant list in X.
Letting you know who wants to talk
When a speaker request arrives, it is read out: "a speaker request has come in from so-and-so". This is separate from the text of replies, and tells you who wants to speak.
When you hear it, check the request in X. If you are mid-sentence, saying "I'll get to it at the break" fits the flow. Note that the bot itself also needs the host to approve its own speaker request before it can talk at all.
Adjusting how much of this you hear
To hear the named announcements, set the reading to "Full" in the web chat's settings, signed in with an account that has control of that Space.
"SE" replaces spoken text with tones and effects, so it will not read the named announcements aloud. "Off" stops sound entirely. Switch between welcoming participants and concentrating on an explanation as the segment requires.
How to stop the reading, and what to check when the bot will not speak, are in the usage guide and the FAQ. Use the announcements alongside the X interface rather than relying on them alone — together they tell you who to bring in next.