Talking to Moe

Four ways in, and they all end in the same place: Moe listening, the notch lit green.

• Hold right ⌥ and talk. Let go when you are done. This one always works and never mishears its name, because it does not listen for one. It needs Accessibility; until that is on, Moe says so wherever you look, on the island when you hover, at the top of the open panel and above the chat box, each with a Fix button, and the warning goes the moment it is on. Settings › Voice › Hold a key to talk changes the key to right ⌘, right ⌃ or fn / 🌐, and the new key works straight away. Right ⌃ is only on some keyboards, and fn works only if macOS is not already using it: set System Settings › Keyboard › Press 🌐 key to Do Nothing, or macOS will act on it too. Everywhere Moe names the key, it names the one you picked. A key held as part of a shortcut doesn't start Moe: hold fn and press Delete, or right ⌘ and Tab, and the press is dropped: nothing is recorded, no picture is taken, and a reply it paused carries on.
• Double-tap the talk key and talk with nothing held, for when your hands are busy or you have a lot to say. The notch breathes and says Listening — tap to send. Pause as long as you like to think: Moe never decides you have finished. Tap the key once and what you said goes to Moe, exactly as if you had held the key and let go; if you said nothing, nothing is sent. Esc drops it without sending, and so does locking the screen or the Mac going to sleep. A tap within a moment of the double-tap (a triple-tap, in effect) is read as "never mind": nothing is sent and Moe stops listening. This is only when Moe is completely at rest. In every other moment — talking, working on an answer, listening after "Hey Moe" or waiting for a follow-up — a double-tap stops everything: what it is saying, the answer it is working on and the microphone. You hear a short low two-note sound and the notch says Stopped for a moment, so you know it worked. While Stopped is showing, another double-tap only stops again.
• ⌃⌥Space to start and stop, hands free. It needs no permission at all. Shortcut to ask, under More in Settings › Voice, offers ⌥Space, ⌘⌥Space and ⌃Space instead, for when another app already owns it. A new shortcut takes effect when Moe restarts.
• Say "Hey Moe": Settings › Voice › Or just say "Hey Moe", off by default. The menu-bar icon has the same switch, Always Listening ("Hey Moe").

A quick tap from rest does nothing you can see: the notch only shows Moe listening once you have held the key for a moment, so the first tap of a double-tap never flashes the microphone on. Nothing you say at the start of a hold is lost; Moe keeps the audio from the instant the key went down.

Esc stops Moe too, when the panel under the notch or the chat window is the window you are typing in. In any other app Esc belongs to that app.

Two switches under Moe and your screen in Settings › Permissions go with the key. Moe sees my screen when I ask is on: a picture of the screen under your cursor, taken as you press the key, goes with your words, so "what do I do here" just works. It is never saved. Move my mouse to what Moe points at is off: Moe only draws a ring around the thing it means, and it never clicks.

Moe starts reading your turn back the moment you stop talking, not the moment it decides you have finished, so when you stay quiet the answer starts sooner. Say anything more and the whole turn is read again; nothing is ever taken from a half-heard sentence.

Hands-free, with nothing held down#

Only one of the four ways needs nothing pressed at all: Say "Hey Moe". Turn it on in Settings › Voice › Or just say "Hey Moe", or from the menu-bar icon, and Moe starts listening from the name alone, with nothing pressed and nothing held.

⌃⌥Space needs nothing held down — it starts and stops a turn with a tap, and needs no permission at all — but a tap is still a key press, not an answer for someone who cannot press one.

If the shortcut does nothing#

• The hold-key method stays silent: it needs Accessibility first. Until you turn that on, Moe says so wherever you'd look for it — hovering the island, at the top of the open panel, above the chat box — each with a Fix button, and the warning clears the moment it's on.
• Another app already answers to the key you picked: right ⌥, right ⌘ and right ⌃ are free on most Macs, but if something else gets there first, pick a different one in Settings › Voice › Hold a key to talk — the new key works straight away, no restart. fn is the one exception: set System Settings › Keyboard › Press 🌐 key to Do Nothing first, or macOS answers to it instead of Moe.
• ⌃⌥Space does nothing, or clashes with another app: under More in Settings › Voice, Shortcut to ask offers ⌥Space, ⌘⌥Space or ⌃Space instead. A changed shortcut takes effect once Moe restarts, not the moment you pick it.
• The key works, then suddenly doesn't — same key, same app: if the moment you held it, it was also partway through a shortcut another app was mid-press on (fn + Delete, or right ⌘ + Tab), the press is dropped on purpose: nothing is recorded, no picture taken, and any reply Moe had paused carries on where it left off. That's Moe staying out of your own shortcut's way, not a miss.

Staying in the conversation#

Keep listening after Moe answers, in Settings › Voice, is on to start with: when Moe finishes answering, it keeps listening for a few seconds so a follow-up needs no key and no name; you just keep talking. Turn it off and Moe answers one question and stops listening. While it is on, two controls sit under it: Waits for chooses 8, 12, 16, 20 or 30 seconds (it starts at 8), and When you go quiet is covered under Ending a conversation. Longer is friendlier in a quiet room and worse in a busy one, because everything said in that window is something Moe has to judge. A sound in the room that starts the microphone and comes to nothing, a television or somebody else talking, does not start the wait over: the window gets a few more seconds, and after the third such sound in a row it closes.

Keep listening after a held-key turn too, under More, decides whether the window opens after a hold-to-talk turn as well, or only after the name and the shortcut. Holding the key is meant as a one-off question, so this is off unless you want otherwise, and it is greyed out while Keep listening is off.

Not everything you say is for Moe#

The hard part of an always-listening assistant is not hearing you; it is working out which of the things it hears were meant for it. In the follow-up window Moe weighs several signals rather than matching a phrase: how long ago it spoke, whether it had just asked you a question, how loud you are compared with the room, whether the voice is the one that started the conversation, and what the words themselves suggest. In English it reads the words closely; in other languages it goes by the name, the timing, the loudness and the voice. If the answer is still unclear it asks the Mac's own on-device judgement (on macOS 26 with Apple Intelligence), and without that it takes a turn that looks like a request and leaves one that does not. A long answer straight after Moe asked you something is taken as the answer, however long it runs, and a turn in your own voice, close to the Mac, is taken even when the words alone leave it unclear. If Moe still decides a long turn was not for it, what you said is not lost: it appears in the conversation marked Not sent to Moe, so you can read it and say it again.

It also knows its own voice. If the microphone picks up Moe's own speech, that is echo and not a new turn; it never answers itself.

Whoever opens a conversation owns it. The moment you say "Hey Moe" or hold the key, Moe takes a reading of that voice and uses it for the rest of the conversation, so a second person talking beside you is not answered, and a guest who opens their own conversation is. A reading that disagrees is not enough on its own to ignore you, though: a turn as loud as you usually are, or one that matches the voice Moe has learned is yours, is judged on its words. Someone else can still join by name: saying "Moe" gets through regardless. This needs no setup, because the voice it compares against is the one that just spoke.

Teach Moe your voice, under More in Settings › Voice, learns what you sound like from a few sentences read aloud. It is optional; with it, "is this the person who started this conversation" becomes a measurement instead of a guess. Press Read one, read the sentence on screen, and after three the row reads Moe knows your voice. Moe keeps learning from the turns you open by name. Nothing is recorded; what is kept is a description of the voice, on this Mac. Forget appears beside it once you have read one: Moe stops using how a voice sounds as a signal, and nothing else changes.

Interrupting#

Brushing the hold key does not cost you the answer. The sound stops the moment the key goes down, which is what you want when you have something to say, and the reply is kept, not thrown away, until what you do next says whether you meant to replace it.

Stop Moe to say something, with the key or by talking over it, and what you say next changes what you asked for rather than replacing it: ask for five things, correct one of them halfway through, and Moe keeps the other four. Withdraw it outright, "never mind", "forget it", "scratch that", and it drops the lot. Amending is the default because it is the one you can undo.

Let me interrupt Moe by talking, off to start with, lets you talk over it without a key. Off means off: while Moe is thinking or answering, nothing you say stops it, restarts it, pauses it or makes it quieter, not even its name, so you can talk to someone else in the room, or to yourself, without losing the answer. The only exception is finishing your own sentence: carry on within about a second and a half of pausing and it is joined to what you asked (see Pausing mid-sentence). The hold key, the shortcut and the buttons on the island still stop or pause it whenever you like. With it on, start speaking while it is answering and, as soon as it hears a word, it fades a little quieter while it listens for the rest. It stops when it hears two words, or one word that means stop: "stop", "wait", "no", "hold on", or its name. A fan, a dish, a cough or a door has no words in it, so Moe carries on at full volume having interrupted nothing, and the same sound does not use up your one interruption. Before it stops, Moe checks that the words are not its own coming back through the speakers and that the voice is not much quieter than yours, as someone across the room or the television would be. A stop word or its name stops it unless it is the very word Moe is saying at that moment. Each reply can be paused this way once; a false alarm does not get a second go until the next turn.

Moe listens for you the whole way through a reply, not just at the moments it goes quiet between sentences. A little of its own voice always comes back through the microphone, so it also listens for a voice clearly louder than that, and when you start talking over it mid-sentence it hears you the same way. In that case it does not fade first: it stops once it has heard your words.

If Moe did stop and it turns out nothing was said to it, it carries on: from the very word it stopped at if the pause was short, or from the start of that sentence if it was more than about three seconds.

The same rule holds while Moe is still thinking about your question: a noise in the room does nothing to it. Only words take the floor.

The same is true while Moe is still thinking: pressing the shortcut again mid-thought puts you back in control. Pressing the shortcut never waits; it stops Moe immediately. Saying "Hey Moe" clearly over an answer does the same, whether or not Let me interrupt Moe by talking is on; any other voice over the answer only stops it when that setting is on.

Pausing what Moe is saying#

Tap the hold key while Moe is talking and it stops there. Tap it again and it carries on from the same word. Nothing is re-said and nothing is skipped. The island says paused — tap again to carry on so you can tell a pause from a problem. It works the same on an answer to something you typed, and on one you are replaying with ▶ in Chat.

The same pause is a button too. On the island, every control sits behind the small ⋯ at the bottom right of the panel: click it and a short row of icons slides out. While Moe is talking they include pause and start over; pause becomes resume, which carries on from the word it stopped at, and start over says it again from the top. Once an answer is finished the same place offers read aloud and copy. Hover any icon to see what it does. In Chat, the answer being said has ⏸ and ↺ beside Moe's name.

A tap never throws an answer away. While Moe is working on it and nothing is being said (before the answer starts, or in the gap between two sentences while it looks something up), a tap changes nothing and the answer comes as usual. While Moe is talking, a tap pauses it and the next tap carries on. If you started talking over Moe and it paused for you, a tap keeps it paused rather than dropping it.

That is a tap, a quick press. Holding the key is still "let me say something": the sound stops the same way, and then your words either replace the answer or, if you say nothing, let it carry on. Double-tapping while Moe is talking or working stops everything outright; double-tapping while it is completely at rest is how you talk hands-free.

Seeing what Moe heard#

The panel shows the whole conversation as a chat: yours on the right, Moe's on the left, and while you talk your words appear in the newest bubble as you say them. The captions icon behind the ⋯ turns on the live transcript: while it is on, the island opens by itself when you start talking, and again when Moe starts to answer, so you can read along without hovering, even with Open by itself set to Never. It fills in white while it is on, and it lasts for that conversation: when the conversation ends it switches itself off.

A long answer. Hovering shows the panel at its usual size; click the notch and the panel opens taller, with the extra room given to the conversation. A long answer stays at its first line while Moe is still writing it, so you can read down from the top instead of being pulled to the newest word, and the scroll bar shows when there is more.

Muting yourself#

In a spoken conversation, the microphone icon on the island mutes you. It is the first icon behind the ⋯ for the whole conversation, while Moe thinks, while it speaks and while it waits for you to say something more. Moe keeps working on what you asked, and keeps talking, but nothing said in the room reaches it: not a word while it works, not a word over the reply, not "Hey Moe".

While you are muted, you can't miss it. The word beside the notch says Muted, a red Unmute button with a crossed-out microphone sits at the left end of the island's controls, away from the ⋯, and when the island folds away after an answer the notch keeps saying Muted, with a crossed-out microphone, on its left side. Nothing there moves when you talk, because nothing you say is reaching Moe.

Only the Unmute button unmutes you. Clicking the notch or the Muted beside it opens the island and leaves you muted, so you can look at what Moe is doing without switching the mute off. If the island has only dropped down because your pointer passed over the notch, the first click on Unmute keeps it open and you stay muted; click Unmute again to unmute. The same switch is in Chat, on the bar above the box: it reads Muted in red, with Unmute beside it.

A mute stays on until you turn it off. It lasts through Moe's answers, through the next question and through the pause after each reply. Moe finishing an answer never unmutes you. While you are muted, Moe doesn't give up on the conversation after a few quiet seconds the way it normally does: it waits for you to unmute. When you do, it listens for your reply with the whole pause to give it. If a muted conversation is left alone for ten minutes, it closes.

Holding the talk key talks through a mute. Moe hears you while the key is down, answers, and you are still muted when you let go, so you can ask one thing without unmuting. A quick tap of the key still pauses and resumes the reply, and never unmutes you. Pressing the shortcut that starts a conversation does unmute you, because that is you asking Moe to listen. When a conversation ends, any mute ends with it, so Moe always listens for its name again.

Pausing mid-sentence#

A breath is not the end of your turn. How long Moe waits depends on how finished you sounded: when you plainly finished a sentence it starts working straight away; when your last words trail off ("and", "so", "um") or you sounded mid-thought, it waits a few seconds of quiet before deciding. Either way it holds its answer back until you have been quiet for a moment longer: about a second and three quarters after a request that could still go on, more when it was less sure, and a little less after a whole answer like "yes", "not yet" or "go ahead", which cannot be the start of something longer. It keeps listening until the moment you would hear its first word. Go on talking inside that window and nothing is lost: the half-answer is thrown away unheard, what you said before the pause and after it are joined into one request, and that is what Moe answers. It has to be you talking, something Moe can write down, even an "uh" while you find the next word; a sound in the room with nothing said in it leaves your question exactly as it was. Stay quiet and the answer arrives exactly when it would have anyway. This applies to turns you start with the name or the shortcut; a held key ends the turn when you let go.

With Let me interrupt Moe by talking off, that window is short: you can finish the sentence you were in the middle of if you carry on within about three seconds of pausing, and after that Moe is working on what you said and nothing you say changes it (see Interrupting).

Talking in a noisy room#

Moe does not wait for the room to go quiet to know you have finished. It listens for when your words stop: a second or so after the last one, it decides whether you sounded finished, even with a fan, music or a television going. If you sounded mid-thought it gives you about three more seconds, or six when your words trail off ("…and then"), and neither a noise in the room nor Moe asking itself the same question again resets that wait; only more words do. When the premium hearing heard no words at all in a sound, Moe reads it on the Mac within a second instead of waiting on the cloud, and the cloud stays on for the next thing you say.

Moe also learns how loud the room is on its own and only treats a sound as speech once it rises clearly above that. And if something starts the microphone but no words ever come, Moe lets it go after a few seconds without reading it or answering it, so a room cannot put words in your mouth.

Ending a conversation#

Say you are done in whatever words you would use, "that's all", "thanks", "I'm good", and Moe closes the conversation. This is Moe's own judgement of what you meant, not a phrase it matches.

With the conversation switch on, if it has been quiet for a few seconds and Moe is not waiting on you, it offers to step out rather than sitting there lit. When it goes quiet chooses how: Ask on the island (a tick and a cross by the notch, the default), Ask out loud, Both, or Just stop listening.

The name#

"Hey Moe" runs entirely on your Mac, on a small model that listens only for those two syllables; the room is not transcribed. It is off by default, and there are honest limits worth knowing:

• It sometimes misses the name, and it sometimes fires on a phrase that sounds like it ("hey Joe", "a moment"). It learned from hundreds of voices in many accents, and from meetings and read speech that never say it, but not from your voice or your room; Teach Moe its name, below, measures it on yours.
• The words after the name are the request, and the name itself is never part of it: when the hearing writes "Hey Moe" as another word ("Emma?", or a bare "Hey,"), Moe drops it rather than answering it.
• When it fires on something that was not the name, the checks above usually stop anything from happening, but the light comes on for a moment.
• Hold-to-talk and the shortcut are unaffected by any of this.

Teach Moe its name. Under the Or just say "Hey Moe" switch in Settings › Voice, once it is on, Teach… opens a card that takes about two minutes. It says what it is for, how long it takes and that the clips stay on this Mac; press Start (or Space) and it shows one phrase at a time in large type — "Hey Moe" twelve times, said normally, quietly, from across the room and so on, then eight sound-alikes such as "Hey Joe". Say each one when it appears. A meter shows the microphone hearing you, the line reads Recording… while you speak and Got it with a tick when the clip is kept, and the card moves on to the next phrase by itself; a row of dots at the top shows how far you are.

Speak as you normally would when you call Moe: the card judges each clip by the same microphone sound the wake word listens to, so a voice loud enough to wake Moe is loud enough to teach it. A clip that is quiet is still kept — the tick reads that one was quiet, with a suggestion to come a little closer next time. If a clip is no good — nothing heard, too short, or so loud the sound broke up — the card says which and what to do about it, and waits: press Try again (or Space) to say it again, or Skip this one. Redo the last one throws away the clip you just recorded and asks for it again. Pause (or Space) turns the microphone off until you press Carry on; from a pause, Finish with these measures what you have so far. Cancel (or Esc) stops at any point and changes nothing: the clips from that run are deleted and Moe listens for its name as it did before.

After the last phrase the microphone turns off and Moe measures how well it hears you, with a bar that fills as it listens back to each clip. Then it shows the result against how it was before — for example Catches 11 of your 12 "Hey Moe"s (was 8) · lets through 1 of 8 sound-alikes (was 2) — with a before-and-after bar for each, and uses the new setting from then on. Done closes the card; Teach again starts over, in a different room, say. The row then reads Moe knows how you say its name, with what was measured and when.

The clips stay in Moe's folder on this Mac; nothing is sent anywhere unless you choose to. Once you have recorded, the card's first and last steps show how many clips are on this Mac, and that menu has two more choices: Send them to the Moe team… opens a new message to hello@moebot.app in your mail app with your clips zipped and attached, so the next version of the wake word learns from real voices in real rooms, and nothing goes until you press Send; Forget my recordings removes the clips from this Mac and puts the wake word back to the setting it shipped with. When a new version of Moe brings a new wake word model, a setting taught to the old one no longer applies and the new model's own is used; press Teach again to fit it to your voice.

What Moe says while it works#

During a long job Moe says what it is doing, "Going through your mail", "Searching your files", and repeats the current one every so often so you know it is still there. Anything it does through a connected service says which service, whether the service is on this Mac or reached through Composio. The switch for this is Tell me what it's doing while it works, under More in Settings › Voice.

The panel keeps one shape for the whole turn. Beside the notch, one word says where things are: Listening, Thinking, Working with a clock, Speaking, Paused, Your turn. Under it is the whole conversation, like any chat: what you said in bubbles on the right (live as you say them, with a green edge), Moe's answers on the left, newest at the bottom; scroll up for the rest of the session. A long message of yours shows its start and More; Moe's answers are always shown whole. While Moe works, its step is the newest row: what it is doing right now, in plain words ("Searching your files", "Working through the page"), and then, in the same place, the answer. Nothing swaps out underneath you. Stop is the square icon behind the ⋯ while Moe works, and it ends the job. The box to type in is always there. Anything you type while Moe is still on the last thing is held, shown in the chat as Sends when Moe finishes, and sent the moment it does; stopping Moe drops it. The steps Moe took are in the Moe window.

Work silently. When a long job is more talk than you want, click the speaker icon behind the ⋯. It shows what is true now: a speaker while Moe is saying what it does, a crossed-out speaker, filled in white, once you have turned that off. For the rest of that job Moe stops saying what it is doing, and when the job is done it says only the answer. The step it is on stays on the row as text, so you can still see it is working. Two things are never silenced: a message waiting on the confirmation sheet for you to press Send, and anything that went wrong. Press it again to hear the updates for the rest of the job. It lasts one job, and the button only appears while a spoken question is being worked on; a typed one is already quiet.

Language#

Moe starts in your Mac's own language, the first language in System Settings › Language & Region that Moe can hear, and in English when none of them is one it knows. Language, under More in Settings › Voice, changes it. Talking to Moe is tuned for English: in another language Moe waits for a moment of silence to know you have finished, and "Hey Moe" is an English phrase, so the shortcut is the reliable way in; a banner at the top of Settings › Voice says so. The fast engine hears English and about two dozen European languages; other languages use a universal engine that downloads once, about 600 MB.

Moe answers in the language you set it to, and stays there. A stray foreign phrase in a turn, which speech recognition sometimes produces, is treated as the mishearing it usually is rather than a reason to switch.

Addresses and numbers#

Numbers arrive written: say "twenty five thousand four hundred dollars" and you get "$25,400". Email and web addresses are written too. "Email jane three at jklumen dot ai" becomes jane3@jklumen.ai; "example dot co dot uk" becomes example.co.uk; letters said one at a time run together. It only happens where the shape is unmistakable, "dot" followed by a real domain ending, so "a dot matrix printer" stays an ordinary sentence.

Moe's own voice#

Moe speaks with Kokoro, which ships inside the app: twenty-eight voices, male and female, British and American, all running on your Mac with nothing sent anywhere. Voice, under Moe's answers in Settings › Voice, is one list of every voice, grouped by where it comes from; a voice you pick is used straight away, and the ▶ beside it plays a sample. The row is there while answers are out loud.

Two more groups are offered in the list. PocketTTS has twenty-six voices and is the only local one that starts speaking before it has finished thinking. It holds half a second of speech before the first word, and longer when your Mac is busy, so it does not run out partway through a sentence and leave gaps. After a minute or more of quiet it warms itself up again while you are still talking, so the first answer after a pause starts as quickly as the rest. It is not in the app: pick one of its voices and a Download row appears under the voice, saying what it costs, 939 MB, once, kept on this Mac; its button fetches it with a bar, a percentage and a Cancel. Until it is here Moe keeps speaking with Kokoro, and the voice you picked is used the moment it lands. Stop a download and what has arrived stays; the next Download carries on from there. Premium is Moe's cloud-rendered voice, streamed back as it speaks. It is part of the Away plans; picking it on the free plan opens Settings › Upgrade.

Each set starts on a voice chosen by ear, and you can change it whenever you like: Kokoro begins on Heart, PocketTTS on Peter, and Premium on Clive, one of three Premium voices.

Premium is given what Moe has to say a passage at a time rather than a sentence at a time, so it speaks the way one person reads a paragraph, and it no longer starts on half a sentence. Settings › Voice shows the voice you are hearing: reopening the page no longer shows Kokoro (or Clive) when Premium and another voice are what is speaking.

If Moe cannot use the voice you picked, it tells you. When Premium is out of reach, no network, a plan that does not include it, this month's allowance spent, Moe speaks with the voice on your Mac instead and a line under the voice says so: which set you chose, which voice you are actually hearing, and why. Your choice is kept, so Moe goes back to it by itself the next time it can.

Premium voice and hearing, on paid plans#

On the Away and Away+ plans Moe can hear and speak through the cloud: sharper ears and a more natural voice than the ones that run on your Mac. Each plan includes a monthly allowance, counted for the whole account rather than per Mac and reset on the 1st:

Moe never holds the keys to those services: it asks for a single-use token each time it opens a connection. While Hey Moe is listening, Moe keeps one fresh token ready, so starting to talk never waits on fetching one. The premium voice gets its connection ready as soon as you start talking, so the first answer after a quiet spell starts as quickly as any other. On a slow connection Moe waits for premium hearing to catch up with what you said rather than deciding too soon that you said nothing, and when your Mac clearly heard you speak it never takes the cloud's silence for an answer. Moe's log records that a connection failed and why, but never the address it was using, the token, or anything you or Moe said. When the allowance is used up, or if anything goes wrong on the way, no network, a slow answer, a service down, Moe carries on with the voice and ears on your Mac without asking, and nothing you said is lost. If your connection drops for a moment, Moe notices within about a second, finishes that answer in the voice on your Mac, picking up where the premium voice stopped rather than starting the sentence again, and shows Connection dropped — using the Mac's voice for a moment once in the conversation. The Mac's voice is the closest match to the one you chose (a British man for Clive), and one answer is never split back and forth between the two: the premium voice comes back at the start of your next answer. Before each answer Moe also checks that its connection to the premium voice is still alive, and replaces it while you are still talking if it is not. If the premium voice runs out of credits, Moe switches to the voice on your Mac at the next sentence and stays with it, and a line in the conversation says so once: The premium voice has run out of credits, so Moe is using its built-in voice for now. Premium hearing covers English and about twenty other languages; in any other language the Mac's own hearing is used.

Premium hearing and voice, under More in Settings › Voice, is on by default and does nothing on the free plan: what decides is the plan, not the switch. Turn it off and, paid or not, your voice never leaves your Mac: Moe hears and speaks here. (What you said, as text, still goes to the AI provider you think with.) The line under it says which is true right now: Cloud voice: on · 41% of this month used, Cloud voice: not part of your plan, Cloud voice: off, this month's allowance is used up · back on a date, or that Moe's cloud could not be reached just now.

Checking the microphone#

Check my microphone and speakers, which appears under Let me interrupt Moe by talking once it is on, says a sentence out loud, listens to how much of it comes back, and tells you whether your speakers are leaking into your microphone. Stay quiet for the eight seconds it takes. If they are leaking, Moe hears itself, and interrupting by voice gets harder; headphones, or the built-in microphone and speakers together, fix it. It also says when a call app has taken echo cancellation away; that is expected, and it comes back when the call ends.

Settings › Voice#

One page for how you talk to Moe and how it answers, in three short sections. A switch that has settings of its own shows them just under it, and only while it is on.

• Starting — Hold a key to talk (right ⌥ to start with) and Or just say "Hey Moe" (off). With "Hey Moe" on, Teach Moe how you say it appears under it.
• The conversation — Keep listening after Moe answers (on), with Waits for and When you go quiet under it; and Let me interrupt Moe by talking (off), which works best in a quiet room on the Mac's own microphone and speakers, with Check my microphone and speakers under it.
• Moe's answers — Answers, Out loud or On screen only; Voice, with ▶ to hear it, while answers are out loud; and How much Moe says, Short, Balanced or Detailed.

More, at the foot of the page, opens everything else in four groups: Starting (Shortcut to ask, Keep listening after a held-key turn too, Language), While Moe works (Say "on it", Tell me what it's doing, Ding when a written answer is ready), Answers (which questions are answered out loud, written answers, the island panel and how long it stays, Typed questions open the Chat window) and Your voice and Premium (Teach Moe your voice, Premium hearing and voice). Each is described in its own section on this page. The switches about your screen are in Settings › Permissions, under Moe and your screen.

Giving Moe a photo or a file#

Press + beside the message box in Chat and choose where it comes from:

• Photos… — from your library, through Apple's own picker, so Moe sees only the photos you choose and needs no access to the rest.
• Files… — anything on your Mac, folders included.
• Take a Screenshot — drag across the part of the screen you mean, the same as ⇧⌘4. Escape cancels.
• Paste from Clipboard — whatever you last copied, if it is a file or an image.
• Recent on this Mac — the three newest things in Downloads and on your Desktop, one click each.

Or drag files anywhere onto the conversation. What you have added waits above the box, with an × to take any of it back, until you send, with words or on its own.

The panel under the notch has the same +, without the recent files, and you can drag a file onto the notch itself: it opens as the file arrives and stays open until you let go. Once something is waiting in its box it stays open, so you can say what the file is for.

Moe opens what you send where it is on your Mac, so a large file costs nothing to send. Up to four photos per message are shown to Moe as pictures. Photos, pastes and screenshots that did not already have a file of their own are kept in an attachments folder inside Moe's own folder.

Links to what Moe read#

Moe is spoken, so it never reads a web address out loud, but the links are there to click. Under an answer that involved the web, Sources lists the pages Moe actually opened, taken from the steps it ran rather than from what it says it did. When Moe cites something in the answer itself it writes a real link, live in Chat and in the panel under the notch; the spoken version says the label and skips the address. Emoji and other symbols in an answer are shown but not spoken, and an arrow is read as a pause. Ask for "the links" and you get exactly these.

Reading the chat#

In Chat, what you said is on the right, in a tinted bubble, and Moe's answers are on the left under its name. Notes that belong to neither of you, like Stopped., sit in the middle. Under an answer, a small chip says what Moe did to get there, 3 steps, and opens into the list; code Moe writes gets its own scrolling box. Next to Moe's name on each answer, ▶ reads it aloud in Moe's voice, even when typed answers are set to stay on screen. While it plays, ▶ becomes ⏸; pressed, it pauses there, and ▶ again carries on from the same word. ↺ starts it over from the top. Tapping the hold key pauses and resumes it too. ⋯ copies the answer, or reads it from the top.

Copying. Right-click any message, yours or Moe's, in Chat or on the island, for Copy message or Copy conversation. To select across several messages, press Select in the bar at the top of Chat: the conversation becomes one page you can drag across, or ⌘A for all of it, and you stay at the latest message. Done goes back. Copy all beside it copies the whole conversation. While a spoken conversation is open, a bar above the box says whose turn it is, Your turn — just talk, no wake word needed, Thinking…, Speaking…, with End conversation beside it.