Most documentation tools make the same quiet assumption: that you’ll hand your procedures to them, and they’ll keep them safe in their cloud. For a lot of software, that’s a perfectly reasonable trade. For documentation, we didn’t think it was.
Because think about which procedures are actually worth writing down. The ones that matter most are often the ones you’re least willing to upload somewhere. How admin credentials get reset. How patient information is handled. The steps that describe your security posture, your regulated workflows, the internal systems you’d never put in a screenshot on someone else’s server. For an IT team or a clinic, “just upload it to our cloud” isn’t a convenience — it’s the exact thing their own policy forbids.
So we built Notavi the other way around.
What “your storage” actually means
Your library — the guides you capture, the screenshots, the version history — lives in a location you choose. That can be a folder right on your machine, or a folder in cloud storage you already pay for and control: OneDrive, iCloud, Dropbox, Proton Drive, SharePoint, whatever you already use. Notavi reads and writes to that folder. It never hosts your library on our servers.
The nice part is that you get the sync you already like for free. If your guides live in a folder your cloud already watches, they show up on your other devices the same way the rest of your files do — under your account, your access rules, your provider. And if you’d rather keep everything on one machine with nothing syncing at all, that works too. The choice is yours because the files are yours.
The honest part about AI
We’re not going to tell you your data never leaves your machine, because that’s not true, and we’d rather be trusted than clever.
Here’s the actual shape of it. Using AI is optional — you can build and edit a guide entirely by hand, and if you do, nothing is ever sent anywhere. When you choose to have Notavi write a guide up for you, the screenshots and steps from that one session are sent to the AI model to generate the write-up, and then they’re done — not stored by us, and never used to train the model. Your library still lives in your storage the whole time. The AI is a thing you point at a session when you want a draft written; it isn’t a pipe that quietly copies your work somewhere.
And nothing is captured in the background, ever. Notavi doesn’t log your keystrokes and doesn’t snap a picture every time you click. A step exists because you pressed a shortcut to capture it — two shortcuts, actually, and you set the keys yourself. There’s no always-on recorder to switch off, because there was never one running.
Why we’d rather do it this way
Keeping things local costs us something. There’s no tidy “share this guide with a link” cloud loop, which is the growth engine a hosted tool gets for free. We decided that was a fair price. The whole reason someone reaches for a tool like this is that their knowledge is valuable — and if it’s valuable, it should stay theirs. Ownership isn’t a feature we bolted on; it’s the shape of the whole thing.
If that’s the trade you’ve been looking for — documentation that writes itself, keeps its own history, and stays in your hands — you can try it for free. Build your own library, keep it wherever you like, and turn on the AI write-up only if and when you want it. We built this for people who care where their knowledge lives. If that’s you, we’d love to hear how it goes.