Files and memory: teach it your business

Maya corrected her sign-off once. New task, no reminders. Warmly, Maya. Ember is smart, but on day one, it doesn't know your business. Today, Maya teaches it hers. And you'll see what makes Murage special: it keeps learning from you. Every bot, in every task. The more you use it, the better it gets. Let's go. Step one: teach it. In a fresh task, Maya clicks the paperclip. And picks her studio policies. Deposits, delivery times, raw files. Then she adds two house rules, in plain words. No weddings on a Sunday. And a travel fee: eighty cents a mile, for every mile past fifty from Portland. Ember reads the sheet, and plays it all back. The policies, and both rules. Now the important bit. Maya says: save all of that to your memory notes. Done. Ember writes it into its memory notes, which load at the start of every new task. Step two: the real test. A brand new task, so Ember starts from scratch. Maya types in a new enquiry. Priya wants a Sunday in June, at a lodge a hundred and fifty miles away. And she asks about photos, and raw files. Ember drafts the reply, and saves it as a file. And here's what it checked. A Sunday? Maya doesn't book Sundays. A hundred and fifty miles? That's an eighty dollar travel fee. A sneak peek in seventy-two hours, the gallery in six weeks, and no raw files. All from what Maya taught it, in a different task. Step three: correct it, once. Maya asks Ember to reply to Tom. The draft is spot on. Except the sign-off. Speak soon? That's not Maya. So she tells it, the way she'd tell anyone. I always sign off with: Warmly, Maya. Ember asks before it changes the saved draft. Allow once. Fixed. And Ember notes it, for every future draft. Now the test. A brand new task, a different enquiry, and no reminders. Warmly, Maya. One correction, and it stuck. And look at the rest. The deposit, the sneak peek, the gallery, no raw files. Everything Maya taught it, in every new task. So how does it remember? In layers. There's the conversation you're in. Each bot's own notebook. What your team and your projects share. Your whole workspace, where every memory is kept, and every change is saved. And the skills you teach it. On top of that, Flux Router adds a memory of its own, switched on by default, so whichever model or bot is working knows what it needs to do its best job for you. Rather keep everything on your machine? Just switch off external memory in Flux. Step four: how it learns. Open Settings, then Memory. Learning is on from the start, and Flux does the learning work. It learns from your chats, and picks up facts and the way you like things done. Automatically. Want to approve each new memory first? Switch on Review before using. And it's not just Ember. Every bot on Maya's team learns from their chats with her. Step five: check what it knows. New task again, and Maya simply asks. What do you remember about how I run my business? Ember lists it back. Her policies, and both house rules. Step six: find your files. Everything Ember writes lands in Files. Pick Ember, and there's every saved copy, newest first. There's the reply to Priya. Saved copies never change, even if the original does. Preview it, tweak it, and send it yourself. Or just ask Ember to send it. That's memory in Murage. Teach it once, correct it once, and it remembers. Every bot learns how you work, and gets better the more you use it. And you stay in charge: ask what it knows, fix it, or tell it to forget. Teach it one thing tonight, and see it get it right tomorrow. Whatever your business, teach it once, correct it once, and every bot starts the next job already knowing how you work. Teach it one thing tonight. Download it free at murage.ai.