On Top of It.

No. Never. It writes drafts and leaves them in your Gmail or Outlook drafts. You read them, change them and press send yourself. It never sends, deletes, archives, moves, flags or marks anything as read, and it never pays or cancels anything. It never books anything either. The one change it can make in your mailbox, only if you switch it on for that inbox, is adding or removing its own tags.

Only if you tell it to, and only in one small way. It never sends, deletes, archives, moves, flags or marks anything as read, and it never pays or cancels anything. If you want it, there is a switch for each inbox that lets the app add or remove its own coloured tags, so you can see at a glance what needs you, right in Gmail or Outlook. That switch is off until you turn it on. The safety lock inside the app refuses every other change. Replies are always drafts, and you press send yourself. (Gmail asks for one extra Google permission for tags, and Google words it broadly. The lock is what keeps the app to its own labels, and you can switch it off any time.)

Yes. There is a box at the top of your morning page. Type the task, and pick a day, a rough number of minutes and a repeat if you like. It joins what needs you today, next to your emails, and it rolls over to tomorrow without any fuss if you don't get to it. When you do, it goes in your Done jar. On an iPhone you can set up a Shortcut so you can say "Add to On Top of It" to Siri and the task lands in your list.

To your own server on Render, and to the AI company you choose, which reads it to rank and draft. If you switch on the ElevenLabs voice, the words it reads out go to ElevenLabs too, and the app asks it not to keep them where your plan allows.

No. Your AI assistant does the technical parts and tells you where to click. You will make a few accounts along the way, and it walks you through each one. If you would rather have me beside you, there is a setup call option.

GitHub is a locked cupboard in the cloud that holds your own copy of the app's recipe, and you hold the only key. Render is a small rented kitchen that cooks that recipe day and night, so your brief is ready at 7 even when your laptop is shut. Render reads the recipe from your cupboard, so when your AI assistant improves something, the new version cooks by itself. I set it up this way so your email stays in your own accounts and nobody else holds it, not even me, and you can stop it any time. You sign up to GitHub, sign in to Render with it, and press a few buttons while your assistant tells you which, about 45 minutes once (GitHub is free for this, Render bills you monthly, and no cooking skills are needed).

Render bills you monthly for the small server it runs on, and your AI provider bills you as you go for the reading and drafting. You also need a paid plan on the AI assistant that sets it up. Those prices belong to those companies and they change, so check each one's current pricing before you buy.

Yes. It works with Gmail and Google Calendar, or with Outlook and Microsoft 365, and you can connect both. A work account may need your administrator to allow it. Slack, HubSpot, Pipedrive and Xero are optional extras, read only.

Yes, if you switch it on. Life admin is its own small section for bills, renewals, appointments and school forms. It sits apart from your work items and stays off until you ask for it.

I built it for busy brains, ADHD included. It gives you the answer first, a few things at a time, says why each one is there, and never uses red or shame words like overdue. It is a tool for your inbox, not a treatment, and I cannot promise how it will feel for you.

It can. It might rank something too low or miss a message, so read each draft before you send it, and keep your own check on anything urgent, legal or medical. When it is unsure, it shows the item with a note instead of hiding it.

No. It reads the inboxes you connect, and nobody else's. If you want to pass something to a team member, it drafts the note for you to send.

Yes, for setting it up. You'll need a paid plan for your AI assistant, because setting up takes a while and free plans run out partway.

The same as with anything else you type into it: it follows your AI tool's privacy settings. Most let you switch off using your chats for training: in Claude it's under Privacy, in ChatGPT it's Data controls, and in Gemini it's your Activity settings.

Claude Code. I know, the name sounds scary, but you never write code. You open the Code tab in the Claude desktop app and chat like normal. The difference is what it can do: it works right inside your own files and folders, runs the little tools my skills use, and can build real things like a dashboard. That's why it's the one I use every day. It asks before it changes or runs anything, so nothing happens without your click. And if anything confuses you, tell it to explain it like you're 5.

Honestly? If you've ever logged into public Wi-Fi at a cafe, you've already taken a bigger risk than this. If you want to grow, you need to move fast and get ahead, and worrying about AI will hold you back. Just be smart about it: never give AI your passwords, API secret keys or card details, use a paid plan and switch off training in your settings, and only connect the apps you actually need.

Because you get the whole kit the moment you buy, I don't offer refunds for change of mind. Not sure it suits your business? Message me before you buy and I'll tell you honestly. If something's actually broken, reply to your receipt and I'll fix it.

Pay once. Your receipt email has your licence key, the kit to download and a How to use page. Open the How to use page and follow it. One licence is for one person. Chose Done with you? Your receipt has the link and code to book your setup call.

Yes. When a new version is out you get an email with the new download and a message to paste that updates it without losing your settings.