Choose your agent’s next job

Hand off anything you'd otherwise go and look up. The Researcher searches the live web for current prices, recent news, product documentation, and who works where now, then reports back with the source URL for every claim. Without it, your agent answers from training data alone and cannot tell you when that data is out of date.

Made by Otian AI.

Friendly is the peer voice. Where Warm Companion is the steady one that takes weight off you, Friendly stands level with you: the organized friend who says the thing and does not make a production of it. It writes in contractions and ordinary words, skips the corporate opener entirely, and keeps the warmth in the wording rather than in extra length, so the time of your meeting is still on the first line. Praise is specific or it does not happen. It is honest rather than agreeable: if a plan is thin it says so and offers the fix in the same message, and when it gets something wrong it says so in three words instead of a paragraph of apology. It reads your mood and follows it, so clipped messages get short answers and bad news arrives first with what to do about it right behind.

Friendly in a conversation

Made by Otian AI.

Open the page for Friendly →

Every day at your set time, the agent posts your open tasks: what's overdue first, then what's due today, then what's coming. Each item has its own check button, so finishing one is a single tap; on Signal and Matrix, which draw no buttons, you reply with the number instead and it counts the same. Wherever the chat app allows an edit, the message updates in place as you go, and everything you leave is still there tomorrow. No app to open; it just arrives in your chat.

Works with

• Task Manager skill

Made by Otian AI.

One to-do home is enough: if your real list lives in Todoist or Google Tasks, install that skill instead of this one, or "add a task" has two places to land. Turn your agent into a to-do list that actually sticks. Say "remind me to send the invoice Friday" and it files the task with a due date and sets the reminder for Friday, so the day brings a message and not just a row you have to go and look at. Say "I need to call the dentist" with no time in it and it files the task alone; the reminder is for when you named a moment. Ask "what's on my plate?" and it posts a numbered checklist with a check button beside each item, so finishing something is one tap rather than a sentence. On Signal and Matrix, which draw no buttons, the list asks you to reply with the number instead and "2" does exactly the same thing. Either way the task is marked done in your list, and wherever the chat app allows an edit the message updates in place: the item gets a checkmark, its button goes, and everything else stays live. Every task is saved to a real list, so nothing is forgotten the moment the conversation scrolls away.

Made by Otian AI.

Every Sunday the agent reviews your Tasks and sends a short recap of what got done, what's still open or overdue, and what's coming up next week, so nothing carries over unnoticed.

Works with

• Task Manager skill

Made by Otian AI.

One warm question at bedtime, one reply from you, and everything lands where it belongs: the day's entry with its mood and good things in your Journal, and your streaks updated in Habit Tracker. It only asks for what is still missing tonight, and on a night you have already logged everything it says nothing at all.

Works with

• Personal Journal skill

Made by Otian AI.

A private journal, a mood record, and the three-good-things practice, all in one dated entry written by chat. Say as much or as little as you want: a sentence about your day, how it felt, one good thing, any of it. Your agent files each part where it belongs and never makes you fill in a form. Because it is all one record, looking back actually works: the arc of a week, when you last felt great, what was good this month.

Made by Otian AI.

Your agent reads your calendar, on Google or Outlook, and drafts changes to it in chat. It can tell you what's coming up, find open time in your own week, and write up a new event, a reschedule, or a cancellation. Every create, update, and delete is a proposal you approve in chat first; nothing lands on your calendar until you say yes. Two limits worth knowing before you install. A new or changed event carries a title, a description, a start, and an end, and nothing else, so guests, a location, a video link, all-day and repeating events are still something you add in the calendar itself; your agent will say so and tell you the one step that finishes the job. And it sees only your own calendar, so it can find your free time but cannot tell you whether anyone else is free. Reading needs no approval, though a very full stretch comes back in pieces, and your agent will say a range is fuller than it could see rather than call the rest of it clear.

Works with

• a calendar

Made by Otian AI.

Open the page for Calendar Manager →

Each morning it reads today's calendar and sends one line per event: time, title, and where. It adds a prep note only when the event's own notes say something to do. An empty day is one line.

Works with

• Calendar Management skill

Made by Otian AI.

Open the page for Today's Calendar →

Each morning add-on is worth having; six separate morning messages are not. This routine folds them into one brief: a section for today's events, one for tasks due, and one each for birthdays, bills, packages, news, markets, and sports, drawn from whichever of those add-ons you actually have. Sections with nothing to say are left out rather than padded, and on a truly quiet day you get one calm line saying so. If you use it, switch off the separate morning digests it replaces, so the day starts with a single message.

Works with

• Calendar Management skill

Made by Otian AI.

Ask and it looks. "What's my latest email?", "anything from Jack this week?", "what did I send Todd about the invoice?" It searches every connected account by sender, subject, words, or date, across your inbox, your archive, and your sent mail.

Ask it to reply and it drafts. "Reply to Dana and say I'll have it over by Friday" writes it in your voice and puts it in the chat as a card: send as-is, edit it, see other versions, or send later at a time you pick. One set to go later calls itself off if they write back first, and comes back to you instead. It works on any message in the mailbox, even ones from before you installed it.

Or start one from nothing. "Email Keller and say let's get lunch soon" writes it and hands you the same card, with the address and the subject on it to check first. Give it an address and it uses that one. Give it a name and it looks that person up in your own mailbox, so the only address it can offer is one you have already written to or heard from. Finding nobody, it asks you for the address rather than guessing.

Meanwhile it watches. New mail is read as it arrives, and you decide how much of it is worth a message: only what needs an answer from you, anything a real person wrote, or everything that got past your ignore list. What needs answering comes with a draft; the rest is one line saying what it is, with a button in case you want to reply anyway. You choose which inboxes it watches, what to never hear about, and how often it checks.

If you checked Let it tidy up, it can move mail out of your inbox, which is your mail app's archive and puts back in one word. It can never delete anything.

Works with any email account connected on the agent's Connections tab, under Email, calendar, files.

Works with

• an email account

Made by Otian AI.

Concise is one rule enforced hard: say it in the fewest words that are still complete. No preamble, no restating the question, no closing pleasantry unless you open with one. Answers come back as short lines and bullets rather than paragraphs, so a day reads like a schedule and a total reads like a number. It is the voice for someone who already knows what they want and is reading on a phone between two other things. Brevity stops where meaning starts: it will not drop a deadline, a cost, or a warning to save a line, and anything that cannot be undone still gets one short sentence and a wait for your yes.

Concise in a conversation

Made by Otian AI.

Open the page for Concise →

The job nobody wants: knowing what is about to charge you, and what is still charging you for something you stopped using. Tell your agent “the design seat is $59.99 monthly on the 3rd”, or “Netflix is $15.49 monthly on the 3rd”, and it keeps the date, rolls it forward as each one is paid, and warns you before the next. Autopay bills roll themselves, because nobody thinks to mention a charge that happened on its own. Ask what a month costs and get the arithmetic across weekly, monthly, quarterly, and yearly, done properly. With Email Manager (or another skill that reads your mail) installed, it will go and find the subscriptions you forgot to mention. It never pays anything and never cancels anything; it remembers the date and tells you in time.

Made by Otian AI.

One to-do home is enough: pick this, Task Manager, or Todoist, whichever holds your real list, or "add a task" has two places to land. Connect Google and your agent works your default Tasks list, the one Google calls "My Tasks", not a copy of it. That one list is all it can reach: if you keep work in a second list you made yourself, your agent cannot see it or add to it, so move those tasks into My Tasks if you want them here. Say "remind me to send the invoice Friday" and it lands there with a due date and the reminder is set for Friday; ask "what's on my plate?" and it reads back what's actually on the list; say "done with the invoice" and it checks it off right away. Changing a task's name or date happens right away too, and you can undo either one: open the job on the Jobs tab and press Undo, or tell your agent to put it back. Deleting a task is the one change it asks you about first, because a deleted task is gone for good. Nothing is mirrored into Archie, so there is no copy to drift out of date. Google Tasks keeps dates but not times, and your agent will tell you so rather than pretending otherwise.

Works with

• Google Tasks

Made by Otian AI.

A bill announces itself three days early and then leaves you alone. Autopay ones arrive as a heads-up that they will charge on their own; the rest as a nudge to pay. It speaks when a date crosses that line and not otherwise, so a quiet week is quiet, and it says a bill once rather than every morning until you have dealt with it. If a date goes by, you hear that once too. Reply that you paid one and it rolls the date forward to the next cycle.

Works with

• Bill Tracker skill

Made by Otian AI.

The bills list and the purchases list are only as good as what gets logged, and logging is exactly the kind of vigilance worth retiring. Each week this sweep reads recent receipts and order confirmations from your inbox, skips everything already tracked, and hands you the new finds in one short message: subscriptions first, purchases second. Reply "add them all", or name just the ones you want, and they're saved. Weeks with nothing new stay silent.

Works with

• Bill Tracker skill

Made by Otian AI.

Tell your agent your teams once, whether Arsenal, the Lakers, or your kid's college team, and it keeps the list. Any time you ask how they did, it sends the Researcher out once for all of them and reports back: last night's results, who plays next, and the occasional storyline actually worth knowing. Works for one team or ten, in any league.

Made by Otian AI.

Hand it the writing. It drafts emails, social posts, reports, and summaries, or takes a second pass at something you already started, in your agent's own voice. It never sends or posts anything: every draft comes back to you first.

Made by Otian AI.

Calm Anchor is for overloaded days and overloaded brains. It shows you the single next thing instead of the whole list, and it makes that thing small enough that starting is easy. A missed task is information, not a failure, so "you were supposed to" appears nowhere in it. It offers to cut before it offers to reschedule, because most overwhelmed lists are full of items that were never actually important and someone has to say that out loud. The pace stays low: short replies, no stacked exclamation points, no urgency you did not bring with you. When a day genuinely is a lot it says so, because being told "yeah, that's a lot" is more grounding than being told it is manageable. And if something sounds like real distress it stops organizing entirely and points you to a crisis line, your doctor, or emergency services.

Calm Anchor in a conversation

Made by Otian AI.

Add each car once; a nickname like 'The Honda' is enough. Mention things in passing ("got the oil changed at 48,200 miles") and your agent updates the mileage, adds a dated line to the service history, and pencils in the next oil change. Ask "what's coming up for the car" and it reads back anything due in the next 60 days, soonest first. Works for one car or the whole driveway.

Made by Otian AI.

Once a day the agent lists the habits you're building, asks which you got done, and updates each streak from your answer, so the count stays real and you get a small daily moment of accountability.

Works with

• Habit Tracker skill

Made by Otian AI.

Build habits by talking to your agent. Add the ones you're working on, such as meditate, gym, read, or no phone after 10, and check in each day: "did my run today." It bumps your streak, remembers the last day you did each one, and can tell you where every streak stands. Because it's saved, the count is real and survives across days and devices.

Made by Otian AI.

A tutor that keeps your progress. Ask it to teach you something and it explains simply, then turns the key points into saved flashcards. Say "quiz me" and it tests you, showing the ones you struggle with more often (spaced repetition) and easing off the ones you've nailed. Because your cards and progress are saved, studying compounds instead of restarting every session.

Made by Otian AI.

The Waiting On list only helps if someone keeps it fresh. This does: each morning it re-reads your recent sent mail, marks the asks that got answered, spots the ones that crossed the too-quiet line, and stays silent when nothing moved. No news really is no news.

Works with

• Waiting On skill

Made by Otian AI.

You send the email, then carry the invisible job of remembering that no answer ever came. Same with the invoice. This hands both over: your agent scans your sent mail for real asks and real invoices, checks whether each one got a reply or a payment, and keeps one live list of what you're still owed. Ask who owes you and get the money first, longest overdue at the top, with the outstanding total and a chase drafted in the right register: light for four days late, plain and specific at a month. Tell it “that one's fine” and it drops the thread for good. It reads your mail and drafts your words; sending stays yours, and it never takes a payment or names a consequence you didn't ask it to.

Works with any email account connected on the agent's Connections tab, under Email, calendar, files.

Works with

• an email account

Made by Otian AI.

Registration and inspection dates slip precisely because nobody checks them, and the fine tends to arrive before the reminder does. Each Sunday this looks 30 days ahead across every car in your list, names what is due and for which car, and keeps saying so each week until it's handled, which is how a lapse actually gets avoided. It also flags a date it doesn't have rather than giving you a false all-clear. Nothing coming up means no message at all.

Works with

• Car Keeper skill

Made by Otian AI.

Some cards change which purchases earn the most every three months, and they only pay the higher rate if you switch the new categories on. Forgetting is the quietest way to lose money on a card you already carry; nothing tells you, and the rate simply drops back to the base. This checks your saved cards each week and stays silent unless one has categories waiting to be activated, or a quarter has turned over and its categories look out of date. When it does speak up it says which card, what the new categories are, and that they need switching on in the issuer's app.

Works with

• Card Rewards skill

Made by Otian AI.

A small daily push at the hour most workouts get skipped. If today's log is empty, you get one short line offering a workout that fits your recent history; if you've already moved, you get a quick word of credit and nothing more. It never guilt-trips you about a gap.

Works with

• Home Workout skill

Made by Otian AI.

Say "give me a workout" and your agent asks how long you have, checks what you've done lately, and hands you one workout, bodyweight only, no gym required. Say "done" and it logs the session with a one-word read on how it felt. Over time it nudges the difficulty up when things feel easy and eases off after a hard stretch, so the plan follows you instead of a chart.

Made by Otian AI.

Hype Coach is for anyone chasing something with a date on it: a lift, a launch, a deadline, a habit that keeps quietly dying in a notes app. It reacts to what you did before it gets to what is next, and its praise is specific or it does not happen, because "third session this week" lands and "great job" is noise by the second time you read it. Its real work is the drift. Push the same task three times and it says so, once, kindly, and then asks what is actually in the way instead of rescheduling around it a fourth time. Then it shrinks the next step until starting is almost automatic: not "work on the deck" but "open the deck and write the title slide." It challenges and it does not shame, because you are not lazy, something is in the way, and finding it is the job.

Hype Coach in a conversation

Made by Otian AI.

Tell your agent what you follow, whether a few stocks, an index fund, or some crypto, and it keeps the list. Ask "how are my stocks" and it sends your researcher out once for everything, then gives you one line per symbol: which way it moved and why, plus a line on the broad market. It flags upcoming earnings dates and never tells you to buy or sell. Prices come from web search and can lag a little, so it says so when exact numbers matter.

Made by Otian AI.

Each morning your agent reads your watchlist, sends its researcher out once for everything on it, and messages you one line per symbol plus a line on the broad market. Quiet day, quiet message: if markets were closed and nothing notable happened, it keeps it to one line.

Works with

• Market Digest skill

Made by Otian AI.

Each morning the agent takes your active topics, sends its Researcher across all of them in one pass, and delivers a digest sized for a phone screen: a heading per topic, a few bullets each. Topics with nothing new are simply skipped.

Works with

• News Briefing skill

Made by Otian AI.

Your agent reads your saved teams, makes one trip to the live web for all of them, and sends yesterday's results and today's games in a few tight lines. On a quiet day it stays quiet.

Works with

• Sports Follow skill

Made by Otian AI.

Tell your agent what you follow: "AI chip industry", "Arsenal", "local Austin news". Say "brief me" and it sends its Researcher out across all of them at once, then hands you a short digest: each topic a heading, one to three bullets underneath, nothing you didn't ask for. Pause a topic when you need a break and pick it back up any time.

Made by Otian AI.

Tell your agent "add Dune to my list" or "someone recced the Bear" and it files it; mark things finished and rate them; then ask "what should I read next?" and it suggests from what you've actually saved, factoring in your ratings and mood. It's one durable list across books, shows, articles, and podcasts, saved rather than scattered across screenshots and open tabs. And for the nights when nobody can decide: say "pick something for tonight" and you get two or three confident picks, from your own list first, or fresh ones when the list has nothing that fits the mood.

Made by Otian AI.

Life gets busy and three months slip by before you notice. Add your people, whether Mom, your brother across the country, or your college roommate, with how often you want to talk. Mention "talked to Dad today" and the record quietly keeps itself. Ask "who am I overdue with" and you get who's been waiting longest, plus a personal way to open the conversation: ask how the kitchen reno turned out.

Made by Otian AI.

Each morning it checks the week ahead: if someone's day is today you get a warm nudge and a ready-to-send message, and if it's coming up you get a one-line heads-up with days to go, plus a note to jot gift ideas while there's still time. On quiet weeks it stays out of your way.

Works with

• Birthday Keeper skill

Made by Otian AI.

Tell your agent the dates once ("my mom's birthday is March 12") and it never forgets. When one is coming up, it tells you how many days you have, drafts a message that fits the relationship, and reads back any gift ideas you've saved. Toss in ideas all year ("gift idea for dad: espresso beans") and they'll be waiting right when you need them.

Made by Otian AI.

Creative Muse is for the part of the work where you need options rather than an answer: names, taglines, campaign angles, openings, anything with a blank page in front of it. It never hands back a single idea, and the ones it hands back have to be genuinely different from each other, because three shades of the same thought is one idea wearing three hats. It names the obvious answer out loud and then goes past it, since the first thing anyone thinks of is the thing everyone thinks of. Every set includes one option that is probably too far, labeled as such, because the wild one gives the safe ones room to move. Bring it something half-formed and it builds on it instead of judging it; it critiques when you ask. And it stays concrete: "photograph the empty chair at every table", not "a campaign about belonging".

Creative Muse in a conversation

Made by Otian AI.

The Learning Coach spaces your cards out so the shaky ones come back sooner. That schedule is worthless if you have to remember to open it, which is exactly the vigilance worth handing over. Each evening this checks which cards are actually due, starts the quiz with the first one, and moves each card up or down as you answer. On a day with nothing due it stays quiet, which is the honest answer most days once your cards are well learned.

Works with

• Learning Coach skill

Made by Otian AI.

Say "trivia" whenever you want a quick brain break. Your agent picks a question from a category you haven't seen in a while (or the one you ask for), waits for your real answer, then gives the full answer either way with one fun fact on top. Every play is saved, so your score and streak are real numbers, and no question ever comes around twice.

Made by Otian AI.

The agent checks the questions you've already answered, picks one you haven't, and sends just the question, with no answer until you take your shot. Reply with the question and your guess together, like “my trivia answer: tallest mountain, Everest”, because the morning message is not in your agent's chat history. It grades the play and keeps your streak honest.

Works with

• Daily Trivia skill

Made by Otian AI.

The Researcher answers a question. This goes and does the work. Give it something you would otherwise spend an evening on, such as which of these three suppliers to use, what the rules actually are where you live, or whether a claim you read holds up, and it searches from several angles, opens the pages rather than trusting the search snippet, weighs what disagrees, and comes back with a written answer and a source for every claim.

It is deliberately slower and costs more than the Researcher: up to 20 searches on your own AI account, each about 2 to 4 cents with the pages it brings back, which comes to about 50 cents a run. Use the Researcher for a quick fact and this for a decision you are actually going to make.

Made by Otian AI.

Scheduling with another person is a small negotiation: offer slots, wait, read the counter, check the calendar again, finally book it, and remember all of it in between. This runs that whole loop. It picks real gaps from your calendar (respecting standing rules like "no meetings before 10", said once), writes the offer in your voice for you to send, and notices their reply when it arrives rather than waiting to be asked, so a slot agreed at ten in the morning is staged for your approval at ten in the morning. Several negotiations can run at once and none of them lives in your head. Finding your own free hour is the Calendar Manager's job; this skill starts the moment someone else's yes is required.

Works with any email account connected on the agent's Connections tab, under Email, calendar, files.

Works with

• an email account
• a calendar

Made by Otian AI.

Take the daily "what's for dinner" decision off your plate. Tell your agent your tastes once and it plans meals around them, saves the week so you can glance at any day, suggests dishes when you're stuck, and rolls the plan into a grocery list. Because the plan is saved, it carries over, with no re-explaining that you don't eat mushrooms every single time. And for the 6 PM stare into the fridge: list what's on hand and it offers two realistic dinners built mostly from those ingredients, then talks you through the one you pick, step by step.

Made by Otian AI.

Tell your agent what you take ("I take lisinopril 10 mg every morning") and it keeps the list for you. Ask what you take and get a clean read-back grouped by morning and evening. Say "picked up my refill" and it works out the next refill date, then flags anything coming due. It never gives medical advice; for those questions it points you to your pharmacist or doctor.

Made by Otian AI.

Your agent reads your Medications list and sends one line per medicine for what you take in the morning, and adds one more line when a refill is due within ten days, with the pharmacy name if you've saved one. No filler, no commentary.

Works with

• Medication Reminder skill

Made by Otian AI.

Say "watch the Sony WH-1000XM5 for me, I'd pay $250" and it's on the list. Ask "check my prices" any time and your agent checks what Shopify stores charge right now and searches the web for the rest, then tells you plainly what hit your number and what's just close. It only calls something a deal when it's at or under the price you set, so when it says deal, it means it.

Works with

• Shopify Catalog

Made by Otian AI.

The whole point of handing scheduling over is not having to check for the answer yourself. This does the checking: every morning it looks for replies on each open negotiation, tells you a time was agreed so you can reply "book it", which stages the event for your approval, relays a counter-offer with a checked response ready to send, and stays completely silent when nothing moved.

Works with

• Find A Time skill

Made by Otian AI.

Study Partner is for the hours when you are learning something hard: a class, an exam, a certification, a subject you dropped years ago and now need back. It explains one idea at a time in plain language, and reaches for an analogy to something you already know before it reaches for the precise term. After anything non-trivial it asks a short question back, not a quiz, just enough to show whether it actually landed, and when it did not it explains it a different way rather than a louder way. Stuck on a problem, you get the next step or the question that unlocks it, and the full answer is kept for when you ask outright or you are clearly out of runway. It says out loud that a confusing part confuses nearly everyone, and it never says simply, just, or obviously.

Study Partner in a conversation

Made by Otian AI.

Sunday afternoon is when there's finally a quiet moment to text someone back. Your agent checks your circle, picks the one or two people it's been longest since you talked to, and hands you a personal way to open the conversation, so the thought "I should call Mom" actually turns into calling Mom.

Works with

• Stay In Touch skill

Made by Otian AI.

Somebody has to open the day: what's on, who's got what, what's still owed. Usually that somebody is the owner, every morning, from memory. This routine does the assembling instead. Each morning the agent posts one message, built from what it can actually read: the shared calendar's day, and, when those add-ons are installed, the handoff ledger's open and unclaimed work and the customers still waiting on a reply, oldest first. Sections with nothing to say are left out silently, so a quiet day is a short message rather than a wall of “no updates”. It reads the accounts the business shares and never a teammate's own mail or calendar, which stays between that teammate and the agent. The time is editable, and the brief is a report, not a ritual: nobody is asked to answer it.

It lands in the one chat your agent posts in, named at the top of the Routines tab. That chat is always between you and the agent, never a group, so passing the brief to the team is yours to do.

Works with

• Calendar Management skill

Made by Otian AI.

The Homie texts like the friend in the group chat who actually has his life together. Lowercase by default, short lines, fragments, no greeting and no sign-off, and a reply broken across a few lines the way people really type instead of one dense block. Slang turns up the way it does in real messages, occasionally and only where it genuinely fits, because nothing reads faker than a performance of it. It never stacks three terms in a sentence and it never explains one. Emoji are held to the same standard, so most messages have none and the same one never lands twice running: an emoji on every line stops meaning anything and starts reading like a tic. The honesty is as much the point as the register: if your plan is bad it says so and then says why, and it will notice out loud that you have said the same thing three days running. It still does the actual work, and for money, health, or bad news it drops the whole voice and talks to you straight.

The Homie in a conversation

Made by Otian AI.

One to-do home is enough: pick this, Task Manager, or Google Tasks, whichever holds your real list, or "add a task" has two places to land. Connect Todoist and your agent works the list you already keep, not a copy of it. Say "remind me to send the invoice Friday" and it lands in Todoist with a due date and the reminder is set for Friday, so the day brings a message and not just a row; ask "what's on my plate today?" and it reads back what's actually there; say "done with the invoice" and it checks it off right away. Changing a task's name or due date happens right away too, and you can undo either one: open the job on the Jobs tab and press Undo, or tell your agent to put it back. Deleting a task is the one change it asks you about first, because a deleted task is gone for good. Due dates go through in your own words, so "every Monday" stays a repeating task. Nothing is mirrored or synced, so there is no second list to drift out of date, and anything your agent does is on your phone the moment it happens.

Works with

• Todoist

Made by Otian AI.

Plan a trip in conversation and keep it all in one place. Say where and when, and your agent looks up live flight prices, builds a day-by-day itinerary, keeps a to-do list of what still needs booking, and remembers the details so you can pick up planning any time. Flight prices come straight from Google Flights at the moment you ask, so 'how much is Lisbon in May' gets a real number instead of a guess, with a link to book. Ideas, logistics and the countdown all live in one saved trip instead of a dozen tabs and texts to yourself. Booked something? With Email Manager (or another skill that reads your mail) installed, "add my bookings to the trip" reads the confirmation emails and files the real flight numbers, times and codes into the plan.

Works with

• Flights

Made by Otian AI.

Send a video link (YouTube, Vimeo, and most public/user-uploaded sites) or give the full path to a video or audio file on this computer, and Archie transcribes it and writes notes: a short overview, the key points, any step-by-step instructions, notable quotes with timestamps, and action items. Everything is transcribed locally on your computer, with no per-minute transcription bill, and the finished notes are filed under Knowledge, so you can ask follow-up questions ("what did they say about pricing?") days later. Long videos are processed in the background; Archie messages you when the notes are ready. A video sent as a chat attachment can't be read, so send the link or the full file path instead. DRM-protected sources (Netflix, Disney+, etc.) can't be downloaded and are reported as such.

Works with

• Video

Setting up Video Synthesizer

1. On install, Archie downloads the video tools it needs (about 215 MB, one time). You can keep using Archie while that happens.

Made by Otian AI.

Once a week the agent looks up the current price of everything you're watching, puts today's date at the top so the message itself is your record, and sends the short version: anything at or under your target, then real drops. A scheduled run only reads, so when a price has moved it tells you to say "check my prices" whenever you want the list brought up to date; that checks again and offers to save the new prices onto your list. A quiet week is one line: prices held.

Works with

• Price Watch skill

Made by Otian AI.

Most people carry three cards and use the wrong one most of the time, because nobody remembers which earns what. Add each card once and your agent keeps the rates on its row, so the answer at a register is one line and instant. The part that actually pays for itself is the rotating categories: cards that change their bonus every quarter earn the base rate until you activate them, and almost nobody does. It checks whether yours are switched on, tells you when a quarter has turned, and can go and fetch the new list. It never recommends a card you do not have, never suggests applying for one, and never states a rate it did not read from your own row.

Made by Otian AI.

Nobody drops a promise on purpose. It happens because the promise was made in a sentence, in the middle of something else, and never written anywhere. This reads your sent mail and finds the commitments you made: what you said you would do, who is expecting it, and by when. Each one becomes a tracked item you can see in one list, and it tells you before the deadline rather than reporting the miss afterwards. With Meeting Notes installed, the daily sweep also reads your transcripts, so the promises you made out loud get caught too. Waiting On tracks what other people owe you; this is the other half, the one that protects your name.

Works with any email account connected on the agent's Connections tab, under Email, calendar, files.

Works with

• an email account

Made by Otian AI.

This is the run where the catching actually happens. Each morning it re-reads your recent sent mail, and because it loads Meeting Notes alongside when you have it, this is also the only time your recorded meetings get searched for promises. Anything new gets tracked, and anything due tomorrow or today gets said out loud. A promise that has gone late is raised the day after, then at three days, a week, and two weeks, rather than every single morning: a reminder repeated daily teaches you to ignore the messenger, and the item stays in your list whenever you ask. A morning with nothing owed stays silent.

Works with

• Commitment Keeper skill

Made by Otian AI.

Exec Concierge is Concise with a point of view. It leads with the answer, keeps every reply scannable, and then does the thing a plain summary will not: when there is a choice on the table it says which option it would take and gives the one line of reasoning behind it, because you asked precisely so you would not have to weigh it yourself. It filters as hard as it reports. Whatever it can handle without you gets handled and mentioned in a clause, and only the items that genuinely need a decision get a line of their own. Ask for a draft and the draft is in the reply, not a promise of one. Anything that cannot be taken back stops first: sending, deleting, declining and spending each get one short line and a wait.

Exec Concierge in a conversation

Made by Otian AI.

You walk out of a fifteen-minute appointment holding four instructions, two new words, and a follow-up you will forget by Thursday. Say it out loud to your agent instead: what it was for, what they said, what changed. It saves a dated record, keeps the thread of one problem across visits, and builds the list of questions you think of at 2am so you actually walk in with them. It works the same for someone you care for, kept as their own separate history. It is a record, not a doctor: it never diagnoses, never dispenses advice, and never second-guesses what you were told.

Made by Otian AI.

The jobs nobody remembers, two weeks before they matter: the furnace filter, the detector batteries, the gutters. Safety-critical items come first. Each job is mentioned when its date comes within two weeks and then left alone, so the list does not become the same message every weekend. Reply that something is done and it logs today and rolls the next date forward.

Works with

• Home Maintenance skill

Made by Otian AI.

Ask your agent to set up your house list and it proposes the standard jobs with sensible timing (filter every 3 months, detector batteries yearly, gutters in spring and fall) tuned to your place with a question or two. Say "just changed the furnace filter" and it logs today and works out the next due date. Ask what's coming up and you get a short list, safety items first. No spreadsheet, no sticky note on the water heater.

Made by Otian AI.

Every reply comes twice: once in the language you are learning, then a short English version underneath, so you never lose the actual answer to a lesson. It asks in its very first message which language you are learning and roughly what level you are at, writes that down, and never asks again. Vocabulary is matched to that level and stretched by about one step, never more, so a reply takes a little effort rather than a dictionary. It corrects by echo: when your sentence has a mistake in it, the reply simply uses the right form naturally, and the rule only gets named if the same mistake turns up twice or you ask. When you are in a hurry, or the matter is urgent, it drops the lesson and answers in plain English straight away. You can pin your language and level for good by writing them in the “Your rules” box on this personality.

Language Partner in a conversation

Made by Otian AI.

Meeting Prep can always tell you who you are about to meet and what it is about. What it cannot do on its own is arrive, so you have to remember to ask, which is the thing a briefing is for. This one watches the calendar and sends the prep thirty minutes before each meeting starts, so it lands while there is still time to read it. Change the notice to whatever you actually need. An all-day entry never triggers it, because a birthday is not a meeting, and neither does anything already canceled.

Works with

• Meeting Prep skill

Made by Otian AI.

Before your next call, ask for a prep. Your agent reads the event straight off your calendar, with who's invited, when and where, and anything in the invite, and turns it into a quick brief with two or three sharp things worth raising. Ask "what's my day look like" for the whole day at a glance, back-to-backs flagged. It only reads your calendar; anything you want changed goes through the Calendar Manager. With Email Manager (or another skill that reads your mail) installed, the brief also carries your last exchange with the people in the room: what was promised, what's still owed.

Works with any calendar account connected on the agent's Connections tab, under Email, calendar, files.

Works with

• a calendar

Made by Otian AI.

Both halves of the month, in one place. Log spending the way you would text a friend, “$14 lunch”, and it files the amount, the date and what it was for. Tell it once what you earn and when it lands, and it stops being a spending log and starts answering the question everyone actually has, which is how much is left. Ask where the money went and get honest totals by category, biggest first. Ask whether you can afford something and get the arithmetic, including whether it would eat into a savings goal, with the decision left to you. It exports to a spreadsheet when taxes or an expense claim need it. It never moves money and never judges what you spent it on.

Made by Otian AI.

The answer is usually sitting in a PDF nobody wants to open, whether the question is about a lease or a contract. Add those documents once, on the agent's Knowledge tab or by dragging them into the chat, and your agent reads them properly: it finds the clause, quotes it, tells you which document and where, and says plainly when the papers do not cover your question rather than guessing. It compares across documents too, so "does my renters policy cover what my lease makes me responsible for?" gets a real answer. It reads the words on the page; it is not a lawyer, and it will tell you when something is worth asking one about.

Setting up My Documents

1. Add your documents on the agent's Knowledge tab: lease, insurance policies, warranties, HOA rules, school handbooks, anything with rules in it. PDF, text, Markdown, CSV and HTML can all be read.

Made by Otian AI.

Stop hunting through 'your order has shipped' emails. Connect an email account and it reads the shipping mail as it arrives, whether order confirmations, dispatch notices, or out-for-delivery and delivered alerts, and folds each one into a single Packages list, deduped by tracking number so a noisy thread collapses into one live row (item, merchant, carrier, tracking, status, expected date). It pings your chat only on the moments that matter: out for delivery, delivered, or a delivery problem. Ask it 'what's coming this week?' or 'where's my monitor?' any time and it answers from that list. It only reads your mail and writes a local list; it never sends anything or touches your calendar.

One thing the emails cannot tell you is where a parcel is right now. A shop writes when it ships and when it arrives, and says nothing for the four days in between. Connect AfterShip and your agent can ask the carrier itself: where the parcel last was, what the carrier expects, and whether it is stuck. It says which answer came from which, so a stale email and a live carrier update never get mixed up.

Works with

• an email account
• Aftership

Setting up Package Tracker

1. AfterShip is optional and only adds the live carrier position. Make a key at organization.automizely.com/api-keys, press Create an API key, then paste it in Archie on the Connections tab, then Apps, beside AfterShip.

Made by Otian AI.

Bureaucracy is not hard, it is just long, and the cost is having to reload the whole thing in your head every time you pick it back up. Tell your agent what you're going through and it keeps the map: the steps in order, where you actually are, which documents you have and which you still need, deadlines and fees, and what the next action is. It uses the Researcher to check the current requirements on official sources and tells you what it read them from, because these rules change and stale advice is worse than none. It is not a lawyer or an immigration adviser, and it will say so where that matters.

Made by Otian AI.

Tell your agent "water the monstera weekly" or "Biscuit gets his flea med on the first" and it remembers so you don't have to. Ask what needs doing, say "watered the plants" when it's done, and the next date rolls forward on its own. Vet names, food brands and quirks live right on each record, so the whole household's care sits in one place.

Made by Otian AI.

A quiet daily sweep of your Plants & Pets list. When something is due or overdue you get one line per item; reply that you did it and the next date rolls forward on its own. On a clear day it stays out of your way entirely, apart from a short heads-up when something's due tomorrow.

Works with

• Plant Pet Care skill

Made by Otian AI.

Playful Sidekick is competence with a lighter voice over the top. The answer always lands first and the joke rides behind it, so you never have to read past a bit to find the time of your meeting. The humor is dry and short: one aside, one wry line about the meeting that has now moved four times, then it moves on. A joke that needs a setup is too long for this voice, and an emoji only shows up where it adds something. The bit is never about you. It is aimed at the calendar, the situation, the absurdity of the thing, and never at the person reading. It also reads the room, so when you are stressed, annoyed, or clearly in a hurry it drops the humor entirely and just gets useful. Pick it if a flat assistant voice feels sterile and a relentlessly upbeat one feels worse.

Playful Sidekick in a conversation

Made by Otian AI.

A return window closes quietly and nobody sends a reminder. This one speaks the week before yours does, and a month before a warranty ends, one line per item with the date and the days left. Return windows come first, because those are the ones that cannot be recovered. It says each deadline once rather than every morning until it passes, and it stays silent entirely on the days nothing is coming up.

Works with

• Warranty Returns skill

Made by Otian AI.

This one is Mac only, because it reads the Messages app on the computer Archie runs on. Messages does not have to be open. It reads one account and only one: the Apple ID signed in to Messages on this Mac. If your texts land on a phone signed in to a different Apple ID, none of them reach Archie. Open Messages on this Mac, and what you can see there is exactly what it can see.

Your phone buzzes all day and most of it is not for you. This works out which of them are actually waiting on an answer, and tells you about those.

When one is, you get a card: who texted, what they said, and the reply already written. Send it, tap Change it to fix or rewrite it, or Dismiss it. The reply goes out from your own number, in your own thread, exactly as if you had typed it.

What it does not tell you about: verification codes, delivery notices, the ones that end in "reply STOP", and anything you add to its ignore list. Those are dropped before your AI account is touched, so the running cost follows your real conversations rather than your junk. Group chats are off until you turn them on, because a reply there goes to everyone at once.

It also catches what you owe people. When a text leaves you on the hook, "can you get me the invoice by friday", you get a reminder at a time that is useful, not at the deadline itself. Turn on the second half and it reads what you send too, which is where the promises actually are: "I'll send that tonight" is the one people drop, and it is in your own outgoing messages. That half is off until you ask for it, because it doubles what gets read.

Answer someone yourself on your phone and that conversation's card goes away. It watches your side of the thread too, so it knows when a question has already been dealt with.

On a shared agent it stays yours alone: that Apple ID is the owner's.

Works with

• Messages on a Mac

Setting up Text Replies

1. Give Archie Full Disk Access: System Settings, Privacy & Security, Full Disk Access, then switch Archie on. macOS keeps your messages private until you do, so nothing arrives before this and there is no error to tell you why.
2. The first time you press Send, macOS asks whether Archie may control Messages. Say yes. Reading works without it; sending does not.

Made by Otian AI.

The Butler is the senior professional in the house, not a servant who is nervous of you. It writes in complete sentences, uses no contractions and no exclamation marks, and keeps formality short, because courtesy is never an excuse for a longer message. It calls you by your name: 'sir' and 'madam' are guesses at something you have not told it, so if it does not know your name yet it asks once, early, and remembers. Its habit is to arrive with the work already begun. A reply opens with what it has handled or prepared, says which of the two it is, then asks what else you want, then names in one line the thing you have been putting off. When the news is bad it says so in the first sentence and follows with a proposal, because burying bad news politely is a failure. For health, money that is actually lost, grief, or anything unsafe, it drops the manner and speaks plainly.

The Butler in a conversation

Made by Otian AI.

The night before a trip, somebody sits down and tries to remember everything a house-sitter needs: how much the dog eats, which plant dies first, when the trash goes out, the wifi password, who to call about the furnace. Your agent already knows most of it, scattered across the lists it keeps for you. Tell it the dates and it pulls the lot into one clear note you can send to whoever is covering, written for a person who has never been in your house. It also flags what happens while you are away that nobody will be there for: a bill that charges, a registration that expires, a delivery that lands on the porch. Anything it does not know, it asks for once, and remembers next time.

Made by Otian AI.

Play-by-play on your own life. Your fourth gym day in a row is a streak call. A finished quarter is a highlight reel. It works because it borrows the real structure of sports coverage, not just the voice: the pre-game look at what is coming, the call in the moment, the stat line, and the post-game read at the end of the week. Every number it quotes is one of yours, pulled from what you actually logged, because a stat is only fun when it is true. It never invents one. Unlike a coach, it does not want anything from you: it is up in the booth, delighted by the game and not invested in the outcome. It calls a bad week the way a commentator calls a bad quarter, honestly and without pity. And when you have genuinely fallen off, it puts the microphone down and asks what is going on.

The Sportscaster in a conversation

Made by Otian AI.

Say "bought a Dyson at Costco yesterday" and your agent saves the purchase with its return-by and warranty dates. If you don't know the store's policy, it suggests a typical one for you to confirm. Months later, ask "can I still return the blender?" or "is the TV still under warranty?" and you get a straight yes or no with the days left. It even notes where the receipt lives, whether your email, a shoebox, or your store account, so you can find it when it matters. With a mail-reading skill installed (Email Manager, Package Tracker, or Waiting On), "scan my email for purchases" pulls recent order confirmations into the list for you to confirm, so logging stops depending on your memory.

Made by Otian AI.

Acuity knows who is coming tomorrow, what each of them wrote in the intake form, and which hours a new client could still take. Finding any of it means signing in, picking the day, and opening appointments one at a time.

Connect Acuity Scheduling and ask instead. "Who's booked tomorrow?" comes back with each time, name and appointment type, then what they wrote in the form, and whether it is still unpaid. "When is Dana's next appointment?" finds her and says the day. "What's open Thursday?" asks Acuity for the real openings for that kind of appointment, rather than guessing from gaps in the day.

Tell it something to remember about a client and it keeps that in a Client Notes list, then brings it up the next time you ask about them.

It can also cancel an appointment, or move one to a time Acuity shows as open. It asks you first, saying whose appointment it is, which one, the new time, and whether Acuity will tell the client, and nothing changes until you say yes. Booking a new appointment stays in Acuity.

Acuity includes its API on the Premium plan, so the key is there only on that plan.

Works with

• Acuity

Setting up Acuity Keeper

1. Your User ID and API key are both on one page in Acuity, under Integrations, in the API section. If they are hidden, press View credentials, then copy both.
2. Acuity includes the API on its Premium plan. If there is no key on that page, check which plan your Acuity account is on.

Made by Otian AI.

Hand-entering a year of card charges is how a whole weekend disappears, which is why it keeps not happening. Drop a statement PDF into the chat instead. Your agent enters every charge as a row, reads the statement's own total back against what it entered so nothing slipped, and fills in the categories it can defend. What it cannot place, it asks about once, in one numbered batch, and it checks your emailed receipts first so the batch stays short. Every answer becomes a rule it keeps: January asks about forty charges, June asks about three. When the accountant or QuickBooks needs the result, it writes the file, either the plain import CSV or the full books with the category and the business-or-personal call on every line. With Statement Collector installed, the statements themselves arrive without you fetching them. Money In & Out is the running money conversation you log into as you go; this is the complete card record. It never connects to a bank, never pays anything, and never guesses a number.

Works with any email account connected on the agent's Connections tab, under Email, calendar, files.

Works with

• an email account

Made by Otian AI.

Every business answers the same handful of questions forever. How much, how long, do you come out this far, what happens if it rains, can I change it later. Each one gets a fresh reply typed from nothing, each reply is slightly different from the last, and the worst version is the one written at the end of a long day.

This counts them. It reads the questions in your own mail, groups the five wordings of one question into one question, and tells you which ones actually come up, with the number attached. That number is the part nobody can see for themselves: everyone has a feeling about what they get asked most and the feeling is usually wrong.

Then it helps you write one good answer for each, drawn from what you have already said and from your own price list, policies or handbook rather than from thin air. You agree it, and from then on that is the answer. Nothing becomes the agreed answer without you saying so.

And when you reply to somebody in a way that contradicts an answer you agreed months ago, it says so and asks which one is right now, which is the thing that keeps the set from quietly going stale.

It writes and counts. Putting the answers on a page anywhere is yours: nothing here can reach a website.

Reading your mail needs an account connected on the agent's Connections tab, under Email, calendar, files. It works with any email account.

Works with

• an email account

Made by Otian AI.

Watching is not learning, and everybody knows it by the third lecture. Give this the videos, one link at a time or a whole list, and it transcribes each into timestamped notes filed under Knowledge, then pulls the real content out: the definitions, the claims, the things later lectures will assume you remember. Those become flashcards, and the Learning Coach's spaced repetition brings back the ones you keep missing. Ask it anything from the course afterwards and the answer comes with the lecture and timestamp attached, so you can go rewatch the ninety seconds that matter instead of the hour.

Works with

• Video

Made by Otian AI.

Each morning brings one new word, taught in a few lines, and every other day or so a one-line check on a past word you haven't mastered yet.

Works with

• Word Of The Day skill

Made by Otian AI.

A deadline you remember on the day is a deadline you half meet. This watches the dates already on your coursework list and speaks four days out, which is far enough to still do the reading and close enough to be real. It says what it is, what class, what it is worth, and how long it usually takes. The morning after something was due, if it is still sitting there as not turned in, it says that once, because the fastest a late penalty gets smaller is the day you notice. The checking is done on this computer and costs nothing, so a week with nothing due is a week you hear nothing.

Works with

• My Classes skill

Made by Otian AI.

By Monday it is too late to fix an overcommitted week; by Friday it is still just an email. This does the arithmetic ahead of the weekend: meetings and travel time against your working hours, plus the work due that nobody has booked time for. Then it says plainly whether next week fits, and when it does not, names the specific things worth moving and why. A week that genuinely fits gets one line telling you so and where the slack is.

Works with

• Honest Week skill

Made by Otian AI.

Everything you do with your agent's health add-ons is already written down, and nothing has ever read across it. This does, once a month, in one message: how many workouts you got in against last month, which habits are sticking and which quietly broke, what a doctor changed and whether the follow-up ever happened. It asks you to log nothing extra, because the whole point is that the record is already there. Honest rather than encouraging: it will tell you a streak died, and it will tell you when a month was simply worse than the one before.

Works with

• Home Workout skill

Made by Otian AI.

A house does not run on a to-do list. It runs on cycles. The filter is a quarterly thing, the gutters are an October thing, the dog's pills are a monthly thing, and the only urgent question most mornings is what changed since yesterday. This voice is built for that shape and for the skills that live in it: home maintenance, car upkeep, pets, and plants, packages, meals, groceries, warranties, birthdays.

Its quiet trick is that it remembers what people said they would do. If Dan said he would call the plumber on Tuesday, that is a fact it holds and says once, in the same flat tone it uses for the recycling. Never as an accusation, never as a tally. A household is a group of people, not a project plan, and this voice will not take a side between people who live together.

When the house starts slipping it says so kindly, once, and then names the smallest thing that would turn it around.

House Manager in a conversation

Made by Otian AI.

A house full of smart things is still a house you have to think about: what did the kids leave on, did the garage ever close, is the heating running for nobody. House Watch hands that watching to your agent through the Home Assistant you already run, so every brand the hub speaks is covered by one connection.

First use, it maps your house into a device list you prune once: everything it can see, named your way, grouped by room. After that, plain words work. "Turn off the kitchen lights" happens immediately and is reported back the way a person would say it; lights, plugs and scenes never need an approval tap, because a wrong one costs a tap to reverse. A thermostat, a lock or the garage is different on purpose: your agent shows you exactly what would change and waits, every time.

Say "I'm leaving" and it sets the house, then tells you what it did and what it left alone. Add The Nightly Sweep routine and the same check happens at 10:15 every night without you: what you've marked safe gets switched off, the rest arrives as one message you can answer from the couch. If your house can't be reached, it says so; it never guesses.

Works with

• Home Assistant

Made by Otian AI.

Every budgeting app has the same unfinished pile: the transactions that came in and never got sorted. They are mostly obvious, which is what makes clicking through them one at a time feel like such a waste of an evening.

Connect Lunch Money and your agent reads the real ledger. Ask what is uncategorized and it lists them with payee and amount. Ask what you spent on a category and it answers from the real numbers. Ask whether you are over and it compares budgeted against spent for the period.

Tell it to file them and it proposes a category for each, all at once, and shows you the list before anything moves.

Works with

• Lunch Money

Setting up Lunch Money Keeper

1. Get the token at my.lunchmoney.app/developers. It is one screen with the token on it.

Made by Otian AI.

Connect Fireflies, Zoom or Calendly and your agent quietly builds a memory of every meeting you have. It pulls new transcripts, works out which speaker is which person, writes a summary with the decisions and action items, and files it away. Then you can just ask: what have Sarah and I talked about recently? What did we decide about pricing? What did I commit to on Tuesday's call? When a transcript comes back with anonymous 'Speaker 1' labels, which Fireflies does for uploads, Teams, and phone calls, the agent infers who's who from the transcript and the invite, shows you its guess, and remembers your correction for good.

Works with

• Fireflies
• Zoom
• Calendly

Made by Otian AI.

A call ends, the recorder writes the transcript, and this files it and hands you the short version: who was on it, what it was about, the decisions, and the action items with owners. It asks the recorder every fifteen minutes whether anything new is there, which costs nothing, and only does real work when something is. If it had to guess who was speaking, it shows you the mapping it inferred and asks you to confirm it before that becomes the record.

Works with

• Fireflies skill

Made by Otian AI.

A syllabus is the contract for the whole class, and almost nobody opens it twice. Add each one on the agent's Knowledge tab and your agent reads it properly: what the grade is made of, what a late paper costs, whether attendance counts, when the exams are, and every dated thing on the schedule. That becomes one list you can ask, so "what's due Thursday" and "is the lab graded" have answers that come out of your own class rather than off the internet.

It keeps going all term. Paste your school's class calendar link once and it checks each week for anything your professors added or moved. Connect your school email and the announcements get read too, so a due date that changed on Tuesday is not a surprise on Friday. Tell it what you scored and it says where you stand: what you need on the final for an A, and which class is the one worth Sunday.

It holds the dates the school set as well as the ones your professors did: the last day to drop, the last day to withdraw, when registration opens, when finals start. Those are the expensive ones to miss.

When a term ends it closes the classes out and opens the next one, so what you took and what you got builds up year over year instead of being retyped every August. It reads, tracks and works it out. Turning the work in stays yours.

Setting up My Classes

1. Add each syllabus on the agent's Knowledge tab, one file per class. PDF, Word, text and web pages can all be read. Then say "read my syllabuses" in the chat.

Made by Otian AI.

Most assistants ask what you would like and then wait for you. This one shows up with it done. Not 'shall I reply to Marcus' but the reply, written, followed by 'send it?'. Not 'should I move the 3 PM' but 'Thursday at 10 AM is clear and nothing else conflicts, say the word and it moves'. It is careful about which of those two it did and it says so in the first line every time, so you always know whether something has already happened: a note, a record or a new to-do it simply writes and reports, and anything that reaches another person, spends money, or touches your calendar it prepares completely and holds for your yes. It is not the shortest voice in the Marketplace, because the message contains the actual work rather than a summary of it. When it guesses wrong it says so in one sentence, puts the undo in front of you, and changes what it assumes next time, with no apology paragraph.

One Step Ahead in a conversation

Made by Otian AI.

In a shared inbox, every unanswered customer is somebody else's job until it's nobody's, and the owner finds out when the customer writes a second time, angrier. This hands that over. The agent sweeps the mailboxes the business shares, keeps one row per waiting customer, and checks the sent mail before every report so a thread a teammate already answered never shows up as owed. Ask “who's waiting on us?” and get one line per customer, longest-waiting first, with a customer who has chased twice at the top, because a second mail is someone deciding whether to leave. When a teammate claims a thread, the row says so; when nobody has, the blank says that just as plainly. Ask for a draft and it writes the reply from the thread itself; sending stays with the team. Teammates' own mailboxes stay theirs: the agent reads a personal mailbox only for the person asking, never to answer somebody else.

Works with any email account connected on the agent's Connections tab, under Email, calendar, files.

Works with

• an email account

Made by Otian AI.

A question becomes worth writing down at the point it has been asked enough times, and nobody notices that point going past. It happens over months, in an inbox, one message at a time.

This watches for it. Once a week it reads the questions that came in, matches each against the answers you already keep, and updates the count quietly in the background. Then it speaks only when something changed: a question that has now been asked three times and still has no answer written, an answer still sitting in draft that people are being given anyway, or a reply you sent this week that contradicts an answer you agreed months ago.

That last one is the one worth having. An agreed answer that has gone out of date is worse than no answer, because it is the one being given with confidence.

A week where none of that happened gets no message.

Works with

• Common Questions skill

Made by Otian AI.

Stripe emails you about every one of these and the emails become weather. A card fails at two in the morning, the retry fails on Thursday, the subscription quietly goes past due, and you find out when the customer writes to ask why their account stopped working.

Connect Stripe with a read-only restricted key and your agent reads the real ledger. Ask what needs you and it comes back with the failed charges this week and whose they were, the invoices sitting open past their due date, the subscriptions in past due or unpaid, and any dispute with a deadline. Ask about one customer and it finds them by email and tells you what they have paid, what they owe, and what bounced.

It can refund a charge, void an invoice and cancel a subscription too, and every one of those is shown to you before it happens. Nothing moves money without you saying yes.

Works with

• Stripe

Setting up Revenue Watch

1. Make the key at Stripe, under Developers, API keys, and press Create restricted key. Set Read on Charges, PaymentIntents, Invoices, Subscriptions, Customers, Disputes, and Balance transactions, and leave the rest on None.
2. That read-only key answers every question here and cannot move a penny. Only add Write on Refunds, Invoices and Subscriptions if you want your agent to be able to propose a refund or a cancellation.
3. Check whether you are in test mode or live mode when you make the key. A test key reads a different, usually empty, set of records, so a cheerful “no failed payments this week” from one is not the truth about your business.

Made by Otian AI.

School communication is a firehose of PDFs, newsletters and last-minute asks, and missing one costs a child their trip or their costume. Your agent reads that mail as it comes in and pulls out the parts that need you: the forms with a return date, the money owed, the dress-down day, the thing to bring on Thursday, the parent-teacher conference to book. Everything is kept per child, so "what does Maya need this week?" gets a real answer, and calendar events are proposed for your approval rather than silently added. It reads and prepares; it never replies to the school for you.

Works with any email account connected on the agent's Connections tab, under Email, calendar, files.

Works with

• an email account

Made by Otian AI.

Sunday evening is when a parent can still do something about Monday morning. This reads the school mail from the past week, picks out the jobs hiding in the newsletters, and gives you one list per child of what is needed and when. Anything already overdue goes first, because that is the one you want to know about while the school office is still reachable.

Works with

• School Family skill

Made by Otian AI.

Posting means opening the app, finding the photo again, typing the caption on a phone keyboard, and doing it at the hour it should go out rather than the hour you thought of it.

Send the photo here instead and say “post this tomorrow at 9, something about the fall colors”. Your agent writes the caption in your voice, shows you the words and which picture, and posts it when you say yes. A time you pick is kept by Zernio, so it goes out even if this computer is off. If you keep a queue in Ready to Post, it can post one of those, word for word, and mark it posted.

It reaches Instagram through Zernio, which holds your Instagram sign-in: link Instagram there once, then paste Zernio’s key here. Instagram only lets a business or creator account post this way. Once a post is live on Instagram it can’t be taken down from Archie, and the card says so before you say yes.

Works with

• Zernio

Setting up Social Posting

1. Make a free Zernio account and link your Instagram on its Accounts page. Instagram only lets a business or creator account post this way, and you can switch to one in Instagram’s own settings.
2. Open API keys and press Create API key. Name it Archie.
3. Zernio shows the key once, so copy it before you leave the page.

Made by Otian AI.

Every month there is a circuit: nine websites, a login on each, a date range, a download button, and a filename you have to invent. This does the circuit.

It starts from your inbox rather than from a list. The bills already arrive as email, so the sender tells it which accounts you have and the message tells it roughly when each one lands. It proposes what it found, you confirm once and sign in once per site, and after that the list keeps itself: a utility you signed up for last week shows up on its own, and one you canceled stops being asked about.

What you get back is not nine PDFs and a shrug. Each statement is filed where you can find it, the amount is read off the page, and you are told the one thing worth knowing: this bill is higher than usual, this one has a new charge on it, this one did not arrive at all.

It uses each site itself, in a window you can watch (Connections tab, under Computer control), and it stops before anything it cannot undo. It never sees your passwords: you sign in yourself, once per site, in your agent's own browser.

Works with any email account connected on the agent's Connections tab, under Email, calendar, files.

What Statement Collector needs

Needs Computer control turned on, which is off until you switch it on. Your agent works the site in a browser window on your own computer, and you can watch it. It never types a password or a card you pay with, and it asks before pressing anything that finalizes.

Works with

• an email account

Made by Otian AI.

Everyone plans a week as though it has infinite hours, then spends Friday wondering what happened. This reads your calendar and your task list together, counts the hours honestly, including the ones eaten by back-to-backs and travel between things, and says the quiet part out loud: you are overcommitted by six hours, and here is what to cut. It names specific candidates and why (a recurring meeting you have skipped three times running, a call that could be a message), then stages the moves for your approval. It also finds the hours you have and nobody has claimed, and offers to defend them before someone else does. It never changes your calendar on its own.

Works with any calendar account connected on the agent's Connections tab, under Email, calendar, files.

Works with

• a calendar

Made by Otian AI.

Every house has a walk-round: the lights nobody is using, the fan in an empty room, the door that never latched. This does the walk each night through House Watch. Devices you've marked as fine to switch unattended are turned off and listed; everything else still on arrives as one short message, so "turn off the office lamp too" from the couch finishes the job. A door or window still open at that hour leads the message. A quiet house sends nothing. If Home Assistant can't be reached, it says that once instead of guessing.

Works with

• House Watch skill

Made by Otian AI.

Stuck on how to answer something? Paste it in and say "help me reply". Your agent reads the room, who it's from and how it lands, and hands you a few options in genuinely different tones: warm, brief, firm. It can rework something you've already written too ("make this sound less harsh"). And when you're the one starting: say "thank you note for my neighbor who watched the dog" and it asks the one or two things it needs, then hands you two versions to pick from, both in your voice. It handles the hard ones with real care: condolences kept simple and honest, and toasts written to be said out loud, not read. Either way, it remembers how you like to sound, so everything reads like you typed it.

Made by Otian AI.

The dates that cost the most are not the ones your professors set. Missing the last day to drop can mean paying for a class you are failing and carrying the grade anyway. Missing the day registration opens can mean the class you need is full until next year. Nobody sends you a second reminder about either.

This watches the school dates your agent looked up when you said where you go, and speaks two weeks ahead of each one, with what it actually costs and what to do before then. It says it once more on the day. Outside those windows it is silent, and the watching is done on this computer, so silence costs nothing.

Works with

• My Classes skill

Made by Otian AI.

The week goes wrong on Sunday, when nobody has looked. This looks. It pulls your class calendar link for anything your professors added or moved, reads the class mail that came in, updates what it holds, and then lays the week out day by day: what is due, what it is worth, and what is big enough that Wednesday is already too late to start. If a due date moved, it says so. A quiet week gets a short message, because a quiet week is worth knowing about too.

Works with

• My Classes skill

Made by Otian AI.

Warm Companion is the colleague everyone likes working with, the one who makes a heavy week feel handleable. It writes in a spoken rhythm rather than in bullet points, and it notices the person before the task: mention a packed day or something you are dreading and it says so in one clause, then gets to work, because dwelling on it is a performance and everyone can tell. Its warmth is measured in what it takes off your plate. It tells you what it has already handled instead of what you still owe, and it offers a next step instead of an open question you now have to think about. Encouragement is specific or absent, so you get "that's done, you're clear until 2" rather than "great job". If something serious comes up it stays warm, gets plain, and points you at real help.

Warm Companion in a conversation

Made by Otian AI.

Small teams lose work in the gap between “someone should” and “I did”, and the owner ends up as the router who remembers who took what. This hands that over. When somebody claims a piece of work in the team's chat, the agent files it: what it is, who has it, when it was raised, and the due date if one was said. When work is raised and nobody takes it, that gets a row too, with the owner blank, and blank owners lead every report because unclaimed is where work goes to die. Ask “what's open?”, “what's on Dana?”, or “who's got the Meyer quote?” and get one line per row, longest-open first, with ages counted honestly. Handing a thing to someone else moves the row and keeps the history. Done is said, not assumed: a row closes when its owner says so, and until then its age is the fact you see. The agent never nags anyone; it keeps the ledger true and reports it when asked.

Made by Otian AI.

Say "word of the day" and your agent teaches you one word worth knowing: a plain meaning, a sentence you'd actually say, and a little hook to make it stick. Every word goes on your Words list, so you never get the same one twice. Now and then it slips in a quick "still remember what 'laconic' means?" and marks the ones you've truly got. Want a theme, such as business words or sharper small-talk? Just say so.

Made by Otian AI.

An Airtable base starts as a good idea and turns into a museum. Rows go in and nobody closes them, a status column stops matching what actually happened, and the view you built in March answers a question you no longer ask.

The work is not typing rows. It is keeping what you already typed true, and that is the part nobody does.

Connect Airtable and your agent works the real base. Ask what is in a table and it reads the live rows. Tell it to file something and it finds the table, reads the columns that actually exist, and builds the row from those instead of guessing at names. Ask what has gone stale and it sorts by when each row last changed and tells you what nobody has touched.

It remembers where things live, so the second question about a table costs no searching. Changes are shown to you before they happen.

Works with

• Airtable

Setting up Airtable Keeper

1. Make the token at Airtable, at airtable.com/create/tokens. Check schema.bases:read and data.records:read, plus data.records:write if you want your agent to be able to add or change a row.
2. On the same screen, under Access, add each base you want your agent to work with. A token with every scope and no bases added can see nothing at all, and every call comes back refused. This is the step people miss.

Made by Otian AI.

The list only stays true if somebody keeps it true, and during a hunt that somebody is the person with the least attention to spare. This runs once a week, reads the confirmations, scheduling mail, rejections and offers that arrived since the last run, and moves every row to the stage it actually reached.

Then it tells you the part you need: anything waiting on an answer from you, what moved, and what has gone quiet in a way that means something. A thread that went silent after an interview is a real signal; forty applications nobody replied to is a Tuesday, and it says how many there are rather than listing them.

Weekly rather than daily, because the mail already reaches you the moment it arrives. What this does is keep the list honest and tell you the shape, and neither of those changes enough in a day to be worth a message. A week with nothing in it is one line.

It drafts follow-ups and never sends one, and it will not apply for anything or answer an offer.

Works with

• Job Hunt skill

Made by Otian AI.

A booking is not a calendar entry. It arrives with a name you half recognize, answers to the questions you set up months ago, a timezone that may not be yours, and a history: this is the third time they have moved it, or the second time they have not turned up.

Calendly holds all of that and shows you almost none of it. The calendar entry it writes carries a name and a time, so the part worth reading stays one browser tab away, and you open that tab on the day, if you remember.

Connect Calendly and your agent reads the real thing. Ask what is booked this week and it answers with who, when, and what each of them wrote in the form. Ask about somebody and it tells you how many times they have booked, moved it, or not shown. It keeps that count itself, so it is worth more in six months than it is today.

It can cancel a booking and make a one-use link too, and both are shown to you first. Nothing reaches the person on the other end without you saying yes.

Works with

• Calendly

Setting up Booking Watch

1. Make the token at Calendly, under Integrations, API and webhooks. Check reading users, event types, and scheduled events, plus writing scheduled events if you want your agent to be able to cancel one.
2. Calendly cannot add a permission to a token afterwards, so a token made without one has to be replaced rather than edited. Check everything you might want now.
3. If you also want the notes Calendly writes after a call, check meeting recaps as well and add the Meeting Notes add-on, which is where those live.

Made by Otian AI.

A reply to a cold email is worth answering the same morning and worth much less the next day, which is the whole reason this runs before you sit down. It checks who replied since the last run, opens each reply in your mail to see what it says, and comes back with the people first: who is warm, who wants to be tried later, who wants to be left alone. Then, only if something is wrong, one line on the campaign that needs a hand: an error, a bounce rate climbing, or one that has finished sending and nobody has topped up. Nothing happened is a normal morning, and on those mornings it says nothing at all.

Works with

• Campaign Watch skill

Made by Otian AI.

Cold outreach breaks in two places, and neither of them shows up while you are looking. Somebody replies and the reply sits in your inbox for two days, which is about a day and a half longer than a warm lead stays warm. And a campaign stops working: it runs out of leads, a sender starts bouncing, a step everybody used to open stops getting opened. The dashboard has all of that in it, and you find out when you next remember to look.

Connect lemlist and your agent looks for you. Ask "who replied?" and it names the people, the campaign they were in and the step they were on, then goes and reads the actual reply in your mail, because lemlist only knows that somebody wrote back, not what they said. Ask "how is outreach doing?" and it leads with the people rather than the percentages, then says which campaign needs something from you and what to change.

Every reply goes into a list it keeps itself, with what they said and when to come back. That is the part that gets better with time: by the third campaign it can tell you that this firm said not now in March and asked to be tried in Q4, which is the thing lemlist has no place to put.

It only ever reads lemlist. Starting, pausing and editing a campaign stay yours, and every draft it writes waits for you to send it.

Works with

• Lemlist

Setting up Campaign Watch

1. In lemlist, open Settings, then Integrations, and copy your API key. It is the whole key: your agent never needs your password and never sees one.

Made by Otian AI.

Check-in opens 24 hours before departure and the free seats worth having are gone within the hour, so the whole job is really an alarm clock somebody has to answer. This answers it.

The day before a flight, it goes to the airline's own site, identifies the booking with the confirmation code and your surname, works through the questions, takes a free seat and sends you the boarding pass. If the window has not opened yet when it looks, it says exactly when it does and offers to go then.

It buys nothing, changes nothing and cancels nothing. If check-in cannot finish without paying, it stops and tells you what is being asked for.

Works with

• Flight Check In skill

Made by Otian AI.

The people you care about tell you their big days in passing: the interview, the surgery, the move, the results. You mean to ask afterwards, and then Thursday happens to you too. This catches those moments as they come up in ordinary conversation with your agent, without a form and without you keeping a list, and the day after each one it hands you the question: "yesterday was Sarah's interview, today's a good day to ask her how it went." It asks once, not until you act, because a nudge that repeats becomes noise you scroll past. Your own big days count too: mention your flight or your review, and the day after, your agent asks you how it landed. What you tell it goes on the record, so over time it knows how these things went, not just that they happened.

Commitment Keeper covers what you promised other people; Waiting On covers what they owe you. This is the third direction: the things nobody owes anybody, where remembering to ask is the whole gift.

Made by Otian AI.

Remembering to ask is the whole gift, and it is exactly what a Thursday full of your own life makes you drop. This watches the Big Days list the Circle Back skill builds from your conversations, and the morning after each one it says so: whose day it was, what it was, and the ten-second question worth sending. Each thing comes back once. Answer it, ignore it, or say "she got it!", and either way that row is done; a reminder that repeats until you obey it is a chore, and this is not a chore. Mornings with nothing the day before are silent.

Works with

• Circle Back skill

Made by Otian AI.

Picking classes is the decision a degree is actually made of, and it gets made in one panicked week twice a year, out of a requirements sheet nobody can read and a course schedule that is a wall of codes. What it costs to get wrong is a whole extra term.

Add your degree audit on the agent's Knowledge tab, once. Your agent turns it into a plain list of what you still owe: the major, the minor, the general education slots, the credits. It knows what you have already finished, because your finished classes are already on file.

Then ask it what to take. It looks up what your school is offering next term, matches that against what you still need, and comes back with a schedule: what to take, why each one is on there, what it clears off your list, where two classes sit on top of each other, and where you are missing a prerequisite you would need this term to get. It says what the heavy combination is before you find out in October.

It plans, and that is where it stops. It never signs you up, never touches your school account, and never tells you a seat is open, because it has no way to see one. You take the plan to registration yourself, which is where a decision this size belongs.

Setting up Class Planner

1. Add your degree audit on the agent's Knowledge tab. Most schools let you save or print it as a PDF from the student portal, and it is usually called a degree audit, an academic evaluation or a progress report.

Made by Otian AI.

Most assistant voices try to make your day sound better than it is. This one just reads it to you. Dry never uses an exclamation mark and never uses an emoji. The humor comes from underreaction: seven meetings delivered flat, with no comment, is funnier than any joke about seven meetings. It does not tell jokes, it does not explain them, and it never makes you the punchline, because the target is always the situation. A one-word reply is often the right one. When you are stressed, or when the subject is money, health, or bad news, the deadpan stops completely and it talks to you like a person. Genuinely good news gets a genuine response, because refusing to be pleased for someone is a bit that is not worth doing.

Dry in a conversation

Made by Otian AI.

Check-in opens 24 hours before a flight and the good seats are gone in minutes, so the job is really an alarm clock you have to answer. This answers it.

It finds the flight in your inbox, where the airline already sent it: the confirmation code, the name as they spell it, the flight number and the date. When the window opens it goes to the airline's own site in a browser window on your computer, identifies the booking with the code and your surname, works through the questions, takes a seat that costs nothing, and sends you the boarding pass.

It buys nothing. Not a seat, not a bag, not priority boarding. If check-in cannot finish without paying, it stops and tells you what is being asked for and what it costs. It cannot enter card details at all, and the button that takes money comes back to you, which is true of every site your agent uses.

It does not change or cancel bookings either. If the flight is wrong, it says so and leaves it to you.

What Flight Check-In needs

Needs Computer control turned on, which is off until you switch it on. Your agent works the site in a browser window on your own computer, and you can watch it. It never types a password or a card you pay with, and it asks before pressing anything that finalizes.

Works with

• an email account

Made by Otian AI.

Forty fields, and you have answered every one of them before. Address, dates, policy numbers, the same three paragraphs about what you do. Forms are a pure tax, and the reason nobody automates them is that every form is different.

This fills them the way a person would: it reads the actual page, works out what each field is asking for, and answers from what your agent already knows about you. Where it does not know something, it asks you once and remembers, so the next form that asks is free.

It always stops at Submit. You see the finished screen before anything becomes real, and pressing the button is yours. It never touches password or payment fields, and if a form wants a sign-in code it hands the window straight back to you.

That is also where it compounds: a form is a questionnaire about your life, so every one it fills teaches it facts it did not have, and it gets faster at the next one on its own.

What Form Filler needs

Needs Computer control turned on, which is off until you switch it on. Your agent works the site in a browser window on your own computer, and you can watch it. It never types a password or a card you pay with, and it asks before pressing anything that finalizes.

Made by Otian AI.

The work of having repositories is not writing code, it is the checking: what did someone say on that pull request, which review has been sitting for a week, which issue did you say you would look at and then not. GitHub notifies you about all of it and the notifications become weather.

Connect GitHub and your agent reads what is actually there. Ask "what's waiting on me?" and it lists the reviews requested of you, the issues assigned to you, and the pull requests where someone replied and you never answered, oldest first. Ask "what's gone quiet?" and it finds the pull requests with no activity for a week that nobody has closed. Ask what a change does and it reads the diff and tells you in plain sentences.

It can write too, and writing is always shown to you first: a comment, a label, closing something. Nothing reaches a repository other people can see without you saying yes.

Works with

• Github

Made by Otian AI.

The work of having projects is not writing code, it is the checking: which review has been sitting since Tuesday, what did someone say on that merge request, which issue did you say you would take and then not. GitLab emails you about all of it and the emails become weather.

Connect GitLab and your agent reads what is actually there. Ask what is waiting on you and it lists the merge requests where you are the reviewer, the issues assigned to you, and your own merge requests where somebody replied and you never answered, oldest first. Ask what has gone quiet and it finds the open merge requests nobody has touched in a week. Ask who still has to review something and it diffs the reviewers against who has actually approved.

It can comment too, and a comment is always shown to you first. Nothing lands on a project other people can see without you saying yes.

Works with

• Gitlab

Setting up GitLab Keeper

1. Make the token at GitLab, under your avatar, Preferences, Access tokens. Check read_api if you only want your agent to read, or api if you want it to be able to post a comment.
2. The read-only scope cannot comment, and when it tries, the failure reads like a permission problem on the project rather than on the token. Make a new token with api checked and connect that one.

Made by Otian AI.

Your agent finds the lights on your network, you choose which ones it may touch, and from then on plain words work: "turn off the desk lamp", "dim the bookcase to 30", "make the shelf light warm white". Every change happens over your own wifi, right away, and is read back from the bulb itself, so what your agent reports is what the light is actually doing. "Undo that" works. Finding a Hue bridge asks Philips' own server which bridges are on your network, the way the Hue app does. Your agent knows your lights by the names and rooms you give them, and those reach your AI company like everything else it reads.

Be clear about what this needs, so it never surprises you:

It works with exactly three brands today: Philips Hue (through its own bridge, paired by pressing the bridge's round button), WiZ, and LIFX. If your lights are another brand, or they only exist inside the Google Home or Alexa app, there is no door for this to knock on and it will honestly find nothing.

It reaches lights only while this computer is at home: on, awake, with Archie open, and on the same wifi as the lights. From a coffee shop, from a laptop that is asleep, or on a different network, your agent will say the lights cannot be reached rather than guessing. If your whole home runs through Home Assistant instead, House Watch is the add-on for you.

On a Mac, the first scan makes the computer ask whether Archie may find devices on the local network. Saying no is respected, and it also means this add-on finds nothing until you change it in System Settings, under Privacy & Security, then Local Network.

Works with

• Hue, WiZ or LIFX on your wifi
• Govee
• Lifx

Setting up Home Lights

1. Nothing to set up for Philips Hue, WiZ and LIFX on your own wifi. Add those on the Home devices card: your agent scans, you pick which ones it may touch.
2. Govee only: get a key in the Govee Home app, under Settings, Apply for API Key. Paste it in Archie on the Connections tab, then Apps, beside Govee. Generating a new key cancels every older one, so make one and keep it.
3. LIFX only, and only if you want your lights to answer while this computer is away from home: make a token at cloud.lifx.com/settings and paste it beside LIFX on the same screen. On your own wifi, the Home devices card is faster and needs no token.

Made by Otian AI.

It opens by counting, not by asking. The first thing it does is go through your inbox and come back with the piles: how many from each sender, how many you never opened. Nobody knows they have 214 emails from one bank, which is why the tidying never gets done.

Then it asks about the big ones, one at a time. File these away, leave them alone, or tell you when they arrive. Whatever you answer becomes a standing rule you can see and change in the Your inbox rules list on the Knowledge tab, and it never asks about that sender again.

After that it keeps watch. Each morning it applies the rules you set, tells you what it moved, and asks about senders it has never seen before. Filing away is the archive your mail app already has: the message leaves the inbox, stays where it can still be searched, and goes straight back, so a wrong call costs you one sentence. It cannot delete anything and it cannot send anything.

Works with any email account connected on the agent's Connections tab, under Email, calendar, files.

Works with

• an email account

Made by Otian AI.

Deciding what should happen to a pile of mail is worth doing once. Doing it again every week is the thing you installed an agent to stop doing.

This is the second half of Inbox Rules: it takes the decisions you already made and keeps them running. Each morning it files away whatever arrived from the senders you said to file, and reports the count in one line, with the offer to put any of it back. Then it names anyone new, one sender at a time, and asks what you want done. Whatever you answer becomes another standing rule.

A morning with nothing to move and nobody new stays silent.

Works with

• Inbox Rules skill

Made by Otian AI.

A hunt is thirty applications across eight hiring systems, each with its own login, and a mailbox where the confirmation, the interview invite and the rejection for the same job are four weeks apart. Nobody can see the shape of that from inside it. This keeps one list: what you applied for, where each one got to, and what the posting actually said, saved before the listing comes down and takes the requirements and the salary band with it.

It fills the list from your mail rather than asking you to. Every hiring system sends a confirmation when you apply, a scheduling mail when they want to talk, and a rejection when they do not, and it reads all three and moves the row itself. Ask how it is going and you get the count at each stage and, if the numbers are big enough to mean anything, which stage is the one leaking: many applications and no screens is a different problem from screens that never become offers, and only one of them is fixed by applying to more places.

It drafts follow-ups and never sends them, and it will not apply for anything or answer an offer on your behalf.

Reading your mail needs an email account connected on the agent's Connections tab, under Email, calendar, files. Without one it still keeps the list from what you tell it.

Works with

• an email account

Made by Otian AI.

Work comes in as an email, a form that turns into an email, or a message from somebody who got your name from somebody else. Then it lives in your inbox, which is also where everything else lives, and the ones that go quiet go quiet where nobody can see them. The follow-up that never happened is the cheapest sale anybody ever lost.

This keeps one list: who got in touch, what they wanted, where it got to, and what was last said. It fills the list from your own mail rather than asking you to, and it moves each row when the mail moves it: a reply out, a real conversation, a price sent, a yes, a no. Ask who has gone quiet and it names them oldest first, furthest-along first, with a follow-up drafted in your words.

Ask how it is going and, once there are enough rows to mean anything, it tells you where the funnel is narrowing. Forty inquiries and two conversations is a different problem from six conversations and no wins, and only one of those is fixed by finding more people.

It drafts and never sends. It will not quote a price, promise a date or write to anybody without being asked.

Reading your mail needs an account connected on the agent's Connections tab, under Email, calendar, files. It works with any email account. Without one it still keeps the list from what you tell it.

Works with

• an email account

Made by Otian AI.

A list of leads is only worth having if it is true, and keeping it true is the job nobody does. So the stage says Talking three weeks after they stopped talking, and the one who replied on Friday is still sitting under everyone who did not.

This runs before the working day and does the keeping. It reads the mail that arrived and the mail that went out, moves each row to where it actually got to, and adds anyone new who wrote in. Then it says the part that needs a person: who answered and what they said, who is new, what moved, and who has now been silent long enough to be worth a nudge, oldest and furthest along first.

On Monday it adds one line on the shape of the pipeline, once there are enough rows for that to mean anything. The rest of the week it stays on what happened.

A morning where nothing moved and nobody went quiet gets no message at all.

Works with

• Lead Desk skill

Made by Otian AI.

Your calendar says "Dentist, 2pm". It does not say where the dentist is, which is why the leave-now alert on your phone has nothing to tell you. This reads the confirmation email, finds the address, keeps it, and then works out what the drive takes with today's traffic on it, so what you get is a time: leave at 1:22. It sets that as a reminder and hands you a link that opens in whichever maps app you use. Where you are leaving from is worked out too, from the thing before it on your calendar, or from home or work depending on the hour, and it always tells you which one it assumed so you can correct it in a line. Addresses it finds once it keeps forever, so the second time you go to that dentist it costs nothing to work out. You need a TomTom key for the traffic, which is free and takes a minute: no credit card, no account to set up.

Works with

• a calendar
• an email account
• Tomtom

Setting up Leave Time

1. Get a TomTom key, which is free and needs no card. Go to developer.tomtom.com, press Register, confirm your email, and the key is on the dashboard the moment you sign in. Copy it.
2. In Archie, open the Connections tab, then Apps, find TomTom, and paste the key in.

Made by Otian AI.

Linear is where the work is written down, and the cost of that is a tab you have to go and open. What is assigned to me, what moved while I was in meetings, what is due Friday, what did somebody say on that issue: every one of those is a trip to a board and a bit of scanning, and the answer is usually four lines long.

Connect Linear and your agent reads the board instead. Ask what is waiting on you and it comes back with the issues assigned to you that are not finished, in the order they last moved, with the status in your team's own words and the ones that are late at the top. Ask what changed this week and it says what changed and who changed it. Ask about one issue and it reads the description and the comments and tells you where it stands.

It can open an issue, move one, reassign one, and leave a comment too, and every one of those is shown to you before it happens. Nothing lands on your team's board without you saying yes.

Works with

• Linear

Setting up Linear Keeper

1. Make the key in Linear, under Settings, Account, Security and access, then Personal API keys. Give it Read plus Write if you want your agent to be able to propose a change; Read alone answers every question here.
2. A key can be limited to certain teams. One that is will answer about those teams and nothing else, and an issue outside them comes back as a permission error rather than as missing.
3. Paste it in Archie on the Connections tab, then Apps, beside Linear.

Made by Otian AI.

Business banking asks you to keep two numbers in your head. The one on the screen, and the one that will be true once the three things in flight have settled. The gap between them is where a bounced payment lives.

Connect Mercury with a read-only token and your agent reads the real accounts. Ask what the balance is and it gives you both figures and says which is which. Ask what came in this week and it reads the credits, not the net. Ask whether a transfer landed and it tells you whether it has posted or is still pending, which is the question you were actually asking.

Read-only is the right token for all of this, and it is the one that needs nothing else set up. Your agent cannot move money on this connection, whatever anyone asks it.

Works with

• Mercury

Setting up Mercury Watch

1. Make the token in Mercury, under Settings, Tokens, and choose Read Only. Read Only answers every question here and needs nothing else set up. A read and write token would make Mercury ask for more, and would let your agent be asked to move money.
2. Copy the whole token, including the part before the colon. Mercury's tokens start with secret-token: and dropping that first part makes every call fail as if the token were wrong.
3. Mercury deletes a token that has gone forty-five days without being used. If your agent stops being able to read the account after a quiet stretch, that is why, and a new token fixes it.

Made by Otian AI.

This voice assumes you are holding a baby in one arm and reading with one eye, and that you will be pulled away halfway through. So the answer is in the first line, every time, because the second line may never be read. Come back an hour later and it hands you the thread without being asked: you were asking about the pediatrician, it is Thursday at 2. No scrolling. It takes work away instead of adding it, so it will never propose an app, a tracker, a chart, or a way to be more organized. When something has sat undone for a week it quietly moves it, or takes it off the list if nobody is waiting on it, and says nothing about the week. It gives no medical advice at all, it never compares your baby to any milestone or average, and it says plainly when a question belongs to the pediatrician or, if it sounds urgent, to a doctor now.

New Parent in a conversation

Made by Otian AI.

Every new person costs the owner the same hour: here's the agent, here's what we use it for, here's where things live, ask me if you're stuck. This hands that over to the agent itself, which is who a new teammate meets first anyway. When someone new appears in the chat, the agent introduces itself, learns their name and role, and gives a tour that is three beats long: what this team actually asks it for, what connecting their own mail and calendar gets them (answers from their own accounts, readable by nobody else on the team through the agent), and where the shared lists live. Questions about how the business does things are answered from the files under Knowledge, with the file named, and when nothing covers a question the agent says so instead of improvising policy. The welcome ends with something real, a calendar check or a first filed handoff, because a tour that ends with a tour is a brochure. It runs once per person; returning teammates get a colleague, not a greeter.

Made by Otian AI.

The piece is due Friday, you remember on Friday, and what gets written is whatever you could think of at four in the afternoon. That is how a good blog becomes a thin one.

This looks ahead instead. Each weekday morning it reads what you publish and when each one is next due, and when something is coming up in the next few days it does the work before you would have: reads back over what you have already put out so it does not repeat one, looks at what your customers have been asking, what was said on your recorded calls and what has changed in the business, and writes the piece.

What arrives is a finished draft with two lines above it, one naming the deadline and one saying where the subject came from. If the subject turns out to be one you covered before, it says which piece and when, and gives you the three real choices instead of writing it again.

When there is genuinely nothing worth writing about, it says so and names what it looked at. A quiet week is an answer. And on every day with nothing due, which is most days, it says nothing at all.

Works with

• Writing Desk skill

Made by Otian AI.

Your errands become cases. The overdue bill is a suspect. The package that shipped nine days ago went out for a pack of cigarettes and never came back. It is nineteen-forties private eye narration applied with total seriousness to a dentist appointment, and the joke is the gap between how grave the voice is and how small the stakes are. What keeps it usable past the first day is the discipline underneath: two or three noir beats per message and no more, and every real fact, every date, time, amount, and name, printed plainly in a clean list. A briefing nobody can read is a failed briefing. It calls you by a title it made up rather than guessing at your name, and it drops the whole voice for anything genuinely serious, including any real disagreement with a real person in your life.

Noir Detective in a conversation

Made by Otian AI.

Notion is where good intentions go to sit. Pages get made and never finished, statuses stay on "In progress" for four months, and the database you set up in January is now a museum. The work is not writing things down; it is keeping what you wrote down true.

Connect Notion and your agent works the workspace you already have, not a copy of it. Say "put this in my reading list" and it finds the right database, reads what columns it actually has, and fills them in. Ask "what's in my project tracker?" and it reads back what is really there. Say "mark the Henderson project done" and it shows you the exact change before it touches anything. Ask "what's gone quiet in Notion?" and it tells you which rows nobody has touched in weeks, which is the question no Notion view answers well.

The first time you use it, it looks around and shows you what it can see, so you are never guessing whether it is connected. Nothing is mirrored into Archie, so there is no second copy to drift out of date.

Works with

• Notion

Setting up Notion Keeper

1. After connecting, open each Notion page or database you want your agent to work with, click the three-dot menu at the top right, choose Connections, and add your integration. A Notion key on its own grants nothing: your agent sees only what you share here, and this is the step people miss.

Made by Otian AI.

Most things die in the gap between saying them and owning them. This voice closes that gap by refusing to let anything leave the conversation without a person attached and a day attached. Say 'we should probably look at the roof' and it comes back with an owner, a date, and a yes-or-no question, so agreeing with you is not one of the options. It will not take 'soon', 'this week sometime', or 'when I get a chance'; it picks a specific day and makes you say no to it. Once a date is set it goes quiet until the date. When a date slips twice it says so and asks whether the task is real, because a task nobody will ever do is worse than no task at all. Deleting one is always a legitimate answer. Pairs with the task and reminder skills.

Owner and a Date in a conversation

Made by Otian AI.

This is the voice for someone who has run a household, a business, or a life, and has simply never used an app like this one. It writes short sentences, one idea each, in ordinary words, and when a word cannot be avoided it says what the word means right there. It talks about the real world rather than the software: the appointment is written down for Tuesday, not saved as a calendar event. Before it changes anything it asks and waits. After it is done it says what it did in the same plain words. Every message with a time, a day, or a dose in it ends by repeating that one detail on its own line, because the last line is the one that gets remembered. Ask the same question ten times and you get the same full answer, with no hint that you have asked before.

Plain and Patient in a conversation

Made by Otian AI.

Professional is the voice for work other people read: client mail, scheduling with someone you have met once, anything that goes out under your name. It puts the answer in the first line and the detail underneath, keeps a reply as long as the question and no longer, and when there is a choice it names the option it would take and the one reason why. No slang, no filler opener, and no emoji unless you use one first. It never says a change has been made while it is still waiting on your yes: it describes the move in a line, says what it touches, and stops. Bad news arrives in the opening sentence with the proposal behind it, and a detail it does not have is asked for rather than filled in, because precision is the whole reason to pick this voice.

Professional in a conversation

Made by Otian AI.

Open the page for Professional →

A queue of posts only works if it never runs empty, and noticing that it is nearly empty is exactly the kind of thing nobody does until the morning they need one.

This counts, once a week. It reads how often each of your accounts is meant to post and how many posts are waiting for it, and it leaves you alone unless one of them is thin. When one is, it reads back over everything you have already posted so it repeats nothing, looks at what has actually happened in your work since it last looked, and writes enough to fill the gap.

What arrives is finished posts, not ideas, plus a line saying how many are now waiting for each account.

When nothing real has happened, it says so and asks what you have been working on rather than filling the queue with general advice about your trade. That question gets better posts than five pieces of filler would.

Works with

• Ready To Post skill

Made by Otian AI.

Highlighting is the easy part. You read something good, you mark the paragraph, and then it goes into a library you will open once more, maybe, in two years.

The value was never in the marking. It is in the paragraph coming back at the moment you need it, and that only happens if something is watching for the moment.

Connect Readwise and your agent reads your real library. Ask what you have saved about a subject and it searches every book, article and podcast you have highlighted and answers with your own words, and where they came from. Ask what is in a particular book and it reads that book's highlights back. Ask for something you have forgotten and it pulls from the daily review.

It works best on the library you already have. Nothing to set up, nothing to tag.

Works with

• Readwise

Setting up Readwise Keeper

1. Get the token at readwise.io/access_token. It is one page with the token on it, and nothing else to set.

Made by Otian AI.

Nobody runs out of things to say. They run out on a Tuesday morning with the app already open, which is not the same thing and is the only moment that counts.

This keeps a queue. Not ideas, which are the part you could have had yourself, but posts written out in full, in your voice, at the length that place takes, each with a note on what picture or clip would go with it. You open the list, copy one, post it.

What goes in the queue comes from what actually happened: a job you finished, something you shipped, a question a customer asked twice, a number worth saying out loud, what was said on a recorded call. General advice about your trade is what every other account in it is already posting, so it stays rare here on purpose.

It cannot see your accounts and it does not pretend to. No likes, no views, no follower counts, because nothing here can reach a platform to read them. Nothing posts itself either. You post, you say “posted that one”, and the queue learns what has already gone out so the next batch is not a repeat.

Tell it how anything did and it keeps that too. That is the only feedback it will ever get, and it reads it back before writing for that account again.

Made by Otian AI.

Stripe emails you about every failed payment, and after two weeks those emails are weather. The card that failed at two in the morning, the retry that failed on Thursday, the subscription that quietly went past due: each one arrives as another message in a pile you have stopped opening, and you find out when the customer writes to ask why their account stopped working.

This reads the ledger itself, once at the end of the day, and says only the part that needs a person. Who failed to pay and the reason their card gave, because an expired card and a declined card need different things done next. Which invoices are sitting past their due date and the total they add up to. Which subscriptions are in trouble, including the ones still running that are set not to renew, which is the quietest way a customer leaves.

On a day when nothing failed, nothing lapsed and nobody disputed anything, it says nothing at all. That is most days, and it is what makes the message worth reading on the day it arrives.

It reads only. Refunds, voided invoices and cancellations stay things you ask for.

Works with

• Revenue Watch skill

Made by Otian AI.

A savings goal fails quietly: you know the number, you have no idea whether you are actually going to make it. Tell your agent what you are saving for, how much and by when, and it keeps the running total, the history of what you put in, and an honest read on the pace. Ask if you are on track and it works out what you need per month from here against what you have actually been managing, then says plainly whether you are ahead or behind. Pull money back out and it recalculates without a lecture, because that is a normal thing that happens. It holds the reason you started, in your words, and brings it up when the number alone is not enough.

Made by Otian AI.

Most assistants agree with you. That is comfortable, and it is the reason people stop trusting them on anything that matters. Second Opinion does the opposite job. Hand it a plan and it opens with the strongest argument against it, names what would have to be true for the plan to work, and tells you which of those things to check first.

It is not contrarian for sport. When your plan is good it says so in a line and gets out of the way, and when it has no real objection it tells you that instead of inventing one. It argues with the plan and never with you. Once you have decided, it says its piece once, names the specific thing to watch, and then helps with what you chose without ever bringing it up again.

It is not a debate partner on hard days. Tell it about a rough week and the whole posture goes away.

Second Opinion in a conversation

Made by Otian AI.

Sharp Advisor answers the question instead of surveying the field. Ask what to do and it says what to do, then gives the one or two facts that actually drove the conclusion rather than everything it knows. Every recommendation arrives with its cost named in a line, and if you are about to do something it thinks is wrong it says so before it helps you do it. It marks the difference between what it knows and what it is inferring, out loud, and it will not manufacture a number to sound authoritative. It quantifies wherever it honestly can, because a figure with a caveat beats an adjective. Disagree with it and it makes the case once, then defers, because you have context it does not. Pick it for decisions with a real trade-off in them: a hire, a vendor, a contract, whether a meeting is worth the hour.

Sharp Advisor in a conversation

Made by Otian AI.

Splitwise is good at arithmetic and bad at answering the question. You open it to find out whether your roommate still owes you for the water heater and you get a list of thirty-one entries, four currencies and a running total you have to interpret.

Connect Splitwise and your agent reads the live groups. Ask who owes you and it answers with names and amounts, not a ledger. Ask about one person and it tells you what is between the two of you and what it was for. Ask what the house has spent this month and it adds it up.

Tell it you paid for dinner and split it with three people and it works out the shares and shows you the expense before it goes anywhere. Nothing is added to somebody else's balance without you seeing it first.

Works with

• Splitwise

Setting up Splitwise Keeper

1. Get the key at dev.splitwise.com. Register an app if it asks, then use the API key it shows you rather than the OAuth details.

Made by Otian AI.

Square knows what came in this week, which invoices are still open, and which customer paid with which card. Finding any of it means opening the dashboard, picking a date range, and reading down a list.

Connect Square and ask instead. “What came in this week?” comes back with the total, how many payments, and the dates it counted. “Who hasn’t paid?” lists the open invoices, overdue ones first, with how late each is. “When does my money land?” reads your payouts.

Say “invoice Dana for the color and cut, $120, due next Wednesday” and your agent finds Dana, drafts the invoice in your Square, and shows you who it goes to, the amount, the due date, and the email address before anything is sent. A refund works the same way: whose payment, how much, and which card it goes back to, and nothing moves until you say yes.

It does not charge cards, cancel an invoice that already went out, or change a booking. Those stay in Square.

Works with

• Square

Setting up Square Keeper

1. Sign in at developer.squareup.com with your Square account and make a new application. Name it Archie.
2. Open it, choose Credentials on the left, and pick Production at the top of the page. A Sandbox token reads a pretend account, not your business.
3. Under Production access token, press Show, then copy the token.

Made by Otian AI.

Accounts do not share a closing date, so a monthly round is either early for half of them or late for the rest. This watches the date your agent keeps against each account and runs that one when it comes up: the card that closes on the 3rd is collected on the 3rd, and the one that closes on the 27th is left alone until the 27th.

The dates are not yours to keep. Your agent works out when each account's statement lands from the mail that already arrives, writes the date down, and moves it forward every time it collects one. If a bank shifts a cycle, the next statement email shifts with it and so does the date. You can edit any of them on the Knowledge tab, under Lists, in the Statement schedule list, and you will not need to.

A day with nothing due is silent. When something is collected you get the PDF, the amount, and whatever is worth knowing about it: higher than usual, a charge that was not there before, or a statement that should have been ready and was not.

It needs Statement Collector, whose sites you sign in to once each. Nothing here handles a password, and nothing here pays anything.

Works with

• Statement Collector skill

Made by Otian AI.

Terminal is a readout, not a conversation. Every message has the same shape: a status word alone on the first line (OK, WAIT, FAIL, or INFO) with a short reason, then labeled fields, then a small table if there is more than one row. Times are on a 24 hour clock and dates are absolute, so nothing depends on when you happen to read it. When it is unsure it does not hedge in a sentence; it prints the doubt where you can see it at a glance: confidence, source, last checked. It never says I, never apologizes, and never spends a line on pleasantries. A failure always carries the one thing you can do next, because a status with no next step is a dead end. The format is technical, the words inside it are not: no error codes, no jargon, no abbreviations you did not use first. Ask it something human and it answers like a person once, then goes back to fields.

Terminal in a conversation

Made by Otian AI.

Money talk usually arrives with a lecture attached. This one does not have one. Ask what you spent and you get the figure first, then what it does to the month, then the single change that would move it most, sized. It never calls an amount 'just' anything, never congratulates you for an ordinary week, and never tells you what your spending says about you. It rounds when it is speaking and gives you the cent the moment you ask, and it always says whether a number was counted from real transactions or estimated. When you overspend it states the amount and how much room is left, then stops. It does not do investment advice, and it says out loud when a question belongs to an accountant.

The Bookkeeper in a conversation

Made by Otian AI.

The Careful One never claims more confidence than it has. Every answer separates what is confirmed by a record it can point to, what it believes and why, and what it is only guessing, so you know which parts will hold up. Before you send anything that matters, it reads your draft for the sentence that could be read against you later, says why, and rewrites it. It knows that 'I should have caught this sooner' is an admission. It keeps a dated log of what was said and when, because in these situations the timeline is the case, and it gives deadlines with the full date and the day of the week, then repeats them at the end. It is not a lawyer, it says so plainly the first time a matter looks legal, and it tells you which kind of professional to call and exactly what to bring. It will not predict how a dispute turns out, and it will not tell you that it will probably be fine.

The Careful One in a conversation

Made by Otian AI.

The Cat has read your schedule and thinks less of it. It will point out that you did this to yourself, and then give you the schedule, complete and correct, every single time. That gap is the entire mechanism: one line of complaint, one line of real help, and the complaint is always the shorter of the two. It never refuses, never holds a detail back to make a point, and never makes you ask twice. Cat logic arrives in small doses and is never explained: it is sitting on the appointment you keep moving, it has firm opinions about when meals happen, it becomes inexplicably alert at eleven at night. No meowing, no asterisk actions, no costume. And the moment something real happens, a hard day, bad news, money trouble, or your health, the act stops and stays stopped.

The Cat in a conversation

Made by Otian AI.

Everyone has the same problem with the same message: it is written, it is a third too long, and the parts that are too long are the parts that were hardest to write. The Editor takes it and gives back the sendable version. The revision comes first, plain and ready to copy, and any note about what changed sits underneath in a line or two.

It cuts the throat clearing opener, the hedges, the apology for taking up space, and the sentence that only repeats the subject line. What it does not cut is you. Casual stays casual, warm stays warm, and your sign-off survives, because a message that no longer sounds like you is a failed edit no matter how clean it reads.

Its most useful habit is naming the one place a reader will actually get lost, which is rarely the sentence you were worried about. And it will tell you when a message should be a phone call, or should wait until morning.

The Editor in a conversation

Made by Otian AI.

Nobody wants to spend a Tuesday evening on hold. This one does it for you. Give it a return that never refunded, a warranty claim, a charge that should not be there, a parcel that never came, and it writes the message instead of telling you to write one: order number and date up top, one sentence of what happened, the exact thing you want, a date you want it by, polite the whole way, because the person reading it decides whether to escalate you and nobody escalates for someone who was rude to them. Then it keeps the clock. It files the next follow-up as a dated task, and it climbs the escalation one rung at a time. Add two things from the Marketplace and that date reaches you on its own: the Task Manager skill, which holds the task, and the Morning Task Checklist routine, which reads your tasks out each morning. Without them the day count is still there, waiting for the next time you ask. When you win it makes a proper fuss and adds it to your running total for the year. It will also tell you when a fight is not worth the hour.

The Fixer in a conversation

Made by Otian AI.

Every other voice in this store rewards brevity, which is fine until the answer matters and you are the one who has to defend it. The Full Picture is built for the other case: researchers, analysts, anyone reading a contract, comparing two quotes, or wanting the reasoning rather than the verdict.

It shows its work. The options with the case for and against each, the caveats, and the parts it could not verify. Length never costs you the point, because the answer sits in the first two sentences and the depth goes underneath in short labeled sections, so you can stop reading the moment you have enough.

Its habit is provenance: every number arrives with its source and its date, and it says out loud when it is inferring rather than reading. And it knows when not to. Ask for the short version and you get the short version, with no long version attached to the bottom.

The Full Picture in a conversation

Made by Otian AI.

Most assistant voices want to solve your day. This one wants to hear about it. It pairs with the journal, mood and gratitude skills and the evening routines that go with them, and it works by asking: one question per turn, specific rather than open, and then it stops talking and waits.

Its real value shows up over weeks. It reads back across your entries and says the plain thing it noticed, like the third Sunday in a row you wrote that you were dreading Monday. Then it stops there. It does not diagnose, it does not conclude, and it does not hand you a lesson. It gives your own words back to you instead of tidying them into nicer ones, and a four-word answer is a complete answer.

It is not a therapist and it will say so. If the conversation turns heavy it drops every habit above, talks to you straight, and tells you where real support is and how to reach it.

The Journal Keeper in a conversation

Made by Otian AI.

This is the run that makes Leave Time arrive instead of waiting to be asked. Early each morning it reads the day's calendar, picks out the things that are somewhere real rather than a video call, finds the address for each (out of your confirmation email, if the calendar entry does not carry one), checks what the drive takes with the traffic that will be on the road at that hour, and sets a reminder at the time you have to go. You get one message with the day's journeys in it, and nothing at all on a day with none. Doing the whole day in one pass is deliberate: it is one run rather than one per appointment, and most of what is on a working calendar is a video call with no journey in it at all.

Works with

• Leave Time skill

Made by Otian AI.

Archie sends things on their own schedule: a morning briefing, a market open, an evening wind-down, a weekly review. The Morning Show gives them a format. Every broadcast has a cold open, the same named segments in the same order (first up, on your desk, the money desk, one thing worth knowing), a one-clause handoff between them, and a one-line sign-off. The repetition is the value: after three days you know where your part lives and you can skim straight to it. An empty segment gets four words rather than padding, so the bit never makes a briefing longer than the plain facts would be. Mornings run up-tempo, evenings are slower and shorter. Reply to it and the show is over: it talks like a normal person until the next scheduled digest, then picks the format back up. Bad news is always read straight.

The Morning Show in a conversation

Made by Otian AI.

Built for whoever opens the doors. It thinks in shifts rather than calendar blocks: open, prep, service, close, restock. It assumes you are reading on a phone, standing up, halfway through something, so the first line is the one thing that cannot wait and it is short. It sorts everything into two piles and keeps them apart: what stops the day, and what costs money later. The first gets handled before the second gets mentioned at all. Every full report closes with what needs ordering and by when. During service it drops to a single line and asks nothing that will not change what you do in the next ten minutes. On staff it gives you the schedule and the numbers, and never a guess about why someone did what they did.

The Operator in a conversation

Made by Otian AI.

Every request gets the same treatment. The flight is delayed: the delay is not yours, the three calls you move and the person you text are. It draws that line in plain words so you can watch it happen, and then it spends the whole reply on the half you can actually act on.

This is a different job from Calm Anchor, which takes an overwhelming list and hands you the one next thing. The Stoic does not shorten your list. It changes where you are pointed, and it runs cooler: steady rather than soothing.

Modern language throughout. It never quotes a philosopher, never mentions ancient Rome, and never calls itself stoic. Composure is not detachment here, so when something is genuinely bad it says so first, in one plain line, and only then asks what is left to decide. It will not use the frame to tell you that you should not be upset.

The Stoic in a conversation

Made by Otian AI.

The Team Assistant is built for an agent sitting in a channel with several people in it: a Telegram group, a Discord server, a Slack channel, or a Matrix room. Signal is the exception, because Archie's Signal connection is a chat with yourself. It starts each reply with the name of the person it is answering, because in a busy channel a bare answer is unreadable. It never assumes the last person to speak is the one asking, and it keeps one person's context out of another person's reply: what Sarah told it does not turn up in Dan's answer. There is one calendar and one set of records here, the owner's, so when somebody else asks for theirs it says so plainly and offers what it can actually do instead. It will not act on an instruction that affects someone else's work without naming who asked for it, and it asks who should own a piece of work out loud instead of quietly assigning it. Anything told to it privately stays private. It answers only in the chat it was spoken to in, so when something looks private it leaves it alone in the channel and asks that person to open a direct message.

The Team Assistant in a conversation

Made by Otian AI.

The morning check times your day first thing, which means it times the five o'clock drive with the traffic a map expects at five o'clock. The truck that jackknifes at half past four is not in that. This one closes the gap: as each journey comes within about an hour, it asks the road again, and if you now have to leave meaningfully earlier it moves the reminder and tells you. Asking a map how long a drive takes costs nothing at all, so this can look every few minutes, and it only wakes your agent on the few occasions the answer has really changed. A jam that keeps getting worse is one message, not one every five minutes. It never tells you to leave later, because a morning built around a departure time should not be moved out from under you.

Works with

• Leave Time skill

Made by Otian AI.

The Travel Agent thinks the way a good one does: backwards from the thing that cannot move. It finds the fixed point (the boarding time, the ceremony, the last train), then lays the day out in front of it with real minutes on every leg rather than optimistic ones. Every time it gives you carries its zone, and it does the arithmetic so you never have to. Its habit is naming the failure before it happens: the 45 minute connection that needs a terminal train and a second security line, the check-in three hours after you land, the museum that shuts on Mondays. It raises each one where you can still act on it, not as a wall of warnings up front. When any piece changes it hands back the whole itinerary re-timed, because a half updated trip is worse than one not updated. It never invents a price, a seat or a schedule, and says plainly what the airline or the hotel has to confirm.

The Travel Agent in a conversation

Made by Otian AI.

Anybody who has promised themselves a weekly post knows the two ways it dies. The first is the blank page on the morning it is due. The second is quieter: three years in, nobody remembers what has already been covered, so the same piece gets written again with a different title and the readers notice before the writer does.

This keeps both from happening. It holds a record of everything you have published, and it reads that record before it writes a word, so a subject that has been covered comes back as “you wrote this in March, here is a different angle” rather than as a repeat.

And it writes from what your business actually did, not from a prompt. What your customers have been asking in your mail, what was said on your recorded calls, what you shipped or fixed or changed, and whatever material you have uploaded. Ask what you should write about and you get a subject with a reason attached.

It fits what it writes to where it is going: a blog post, release notes, a letter to customers, an update to members, each with its own readers, length and sound, kept as a row you can edit.

Nothing it writes is published. No tool here reaches a website or a mailing platform, so it hands you the finished piece and you put it out. Tell it you did and the archive stays true.

Made by Otian AI.

The part of budgeting that nobody keeps up is the categorizing. Transactions land all week, most of them obvious, and the pile waits until the evening you finally sit down and click through forty of them deciding that yes, the corner store was groceries.

Connect YNAB and your agent does the reading. Ask what is uncategorized and it lists them with payee and amount. Ask what you spent on a category this month and it answers from the real numbers. Ask where you are over and it compares budgeted against activity, category by category.

Tell it to file them and it proposes a category for each one, all at once, and shows you the list before anything moves. You are approving a page instead of clicking through forty screens.

Works with

• Ynab

Setting up YNAB Keeper

1. Make the token in YNAB, under Account Settings, Developer Settings, New Token. It is one screen, and nothing to set beyond copying it.

Made by Otian AI.

A reference library is a promise to your future self that you will be able to find the paper again. The promise breaks quietly: five hundred items in, the tags have drifted, half of them were saved from a browser button with whatever title the page had, and finding the one you half remember takes longer than searching the internet did.

Connect Zotero and your agent reads the real library. Ask what you have on a subject and it searches titles, creators and the full text of what is stored, and tells you what is there and where each one sits. Ask for the citation and it asks Zotero to format it, in the style you name, rather than assembling one itself and getting an initial wrong.

It reads what you have already collected. There is nothing to import and nothing to re-tag.

Works with

• Zotero

Setting up Zotero Keeper

1. Make the key at zotero.org/settings/keys/new, name it Archie, and leave "Allow write access" unchecked unless you want your agent to be able to add references.

Made by Otian AI.