Why "Talpatra"
For roughly 2,000 years, a large share of what India wrote down — scripture, medicine, astronomy, law, poetry — was preserved on talpatra: dried, treated palm-leaf pages, inscribed with a stylus and rubbed with ink. No servers. No publisher. No subscription. A leaf, a blade, and whoever was willing to copy it out again before it wore through. Some of those manuscripts outlasted the empires that produced them, not because the format was fancy, but because it needed nothing else to keep working. The leaf itself was the whole system.
That's the line I wanted to draw with this app. Your scans live on your device. Nothing else is required for them to keep existing — no account to stay logged into, no server that has to stay up, no company that has to stay in business. Talpatra Scan is named after a storage medium that didn't need permission from anyone to keep working. Neither does this one.
This is one person, not a company
Talpatra Scan is built and maintained by a single developer — me. There's no team behind it, no support queue, no investor deck. If you email the address at the bottom of this page, I'm the one who reads it and the one who wrote the code you're asking about. I'm saying this plainly because most product "about" pages dress up a solo project as something bigger, and I'd rather you know exactly what you're trusting.
Why build another scanner app
I scan a lot of documents for personal projects, and I kept running into the same wall: every "free" scanner app on my phone was free in the way that should make you suspicious. Either your documents got uploaded to someone's cloud as part of the deal — for OCR, for "backup," for a feature you didn't ask to trade your data for — or the app was free of charge but not free of a tracking SDK and an ad SDK running in the background regardless of what the store listing claimed. Reading a permission list and a privacy label after the fact isn't the same as knowing, and I wanted to actually know.
So the constraint came first, before a single screen was designed: the app would never hold the Android INTERNET permission. Not "we won't use it," not a setting you have to find and turn off — the permission itself would not exist in the build. Everything else — the boundary detection, the book dewarp, the on-device OCR, the export pipeline — got built inside that constraint, not around it.
We don't know what you scan, and we designed it so we can't.
What "designed so we can't" actually means
It means the no-network rule isn't a promise sitting next to code that could break it — it's enforced by the build itself. The app's Gradle build includes a task that inspects the final merged Android manifest and fails the build outright if theINTERNET permission is present in it, from any dependency, at any point. An app without that permission cannot open a network socket. It cannot phone home, cannot silently sync, cannot send a crash report without you explicitly reviewing and sending an email draft yourself. There's no telemetry SDK to disable because there's no telemetry SDK, full stop. If you want the complete accounting of what does and doesn't leave your device, that's laid out on theprivacy page.
What's next
Right now Talpatra Scan is pre-launch: no Play Store listing yet, a signed APK on GitHub Releases is the way to get it, and it's free with no ads. The near-term plan is a public launch — getting the Play Store listing live and the app in front of more people. Further out, roughly three to six months after that launch, I plan to introduce an optional one-time lifetime "Pro" unlock covering some extras — things like additional book-dewarp controls, batch export, and export profiles. It isn't available yet, and there's no price set. Anyone using the app before that unlock exists will keep whatever they already have, permanently — nothing you can already do today gets taken away or put behind a paywall later.
Questions, bug reports, or just want to say the book dewarp worked on something difficult — emailhello@talpatra.com. I read everything that comes in.