Building in public
Where does your data live? What runs on our servers and what stays on your computer
A small part of Mimir runs on our servers; the rest runs on your computer. This post explains which data lives where.
Mimir team
An e-commerce catalog is a pile of files full of price lists and products that haven't launched yet. So a brand's first question, quite reasonably, is where that data lives.
On our servers: the crawl that should never stop
The site crawl has to run every day, at any hour. That's why it runs on our servers, and why the titles, descriptions, status codes, and speed measurements of your site's public pages are stored there too. The crawl keeps going when you shut down your computer. Beyond that, the only thing on our servers is your account: your email address, your plan and subscription details, and your connected devices.
On your computer: catalog, measurements, and memory
Your product files, the copy Mimir writes, drafts awaiting approval, AI visibility measurements, research results, and what Mimir has learned about your project (its project memory and knowledge graph) all stay on your computer. You can copy, move, or delete these files.
The model runs on your account
Requests to the model go out from your own Claude account. You're billed directly; Mimir doesn't put a separate account in between.
Your phone is just a remote control
Munin is the assistant that sits in the notch on your Mac; on your phone, it appears on the lock screen. When a draft is waiting for your approval, you can approve or reject it from your phone. On the Mac, the A and R keys are all it takes. The phone connects to Mimir on your computer.