How we build nopa, part 1

A thirty-second task turns into a quest

I needed to merge two PDFs. A thirty-second job: upload, click, download. The first site from search asked me to create an account. The second merged the files but would only let me download them with a watermark — or a subscription. The third had three Download buttons, and you find out which one is real only after you click.

That quest repeats with every small task: make a QR code, format JSON, compress a photo, remove a background. There are plenty of tools, but something always stands between you and the result — signup, a paywall, an ad shaped like a button. And meanwhile your file has already been uploaded to someone else’s server, and nobody will tell you what happens to it there.

What we want nopa to be

The goal is simple and ambitious: a complete set of simple solutions for everyday tasks — PDFs, images, QR codes, text, notes, boards — in one place, with zero barriers. One page, one task. Open it, do the thing, take the result. From your phone, in half a minute, with no “start your free trial”.

And the second thing, the reason this exists at all: your data should not go anywhere. Not “we promise to delete your files in two hours” — different: your files never leave your browser. nopa has no server that could receive them; the only place we can process your photos and documents is your own device.

The practices we hold ourselves to

That goal turned into rules we check on every tool before it ships:

• The result comes before everything else. No “sign up to download” — not ever.
• Free means free. No asterisks, no watermarks, no premium features taking half the screen.
• Three taps to the result on a phone — we count them literally.
• Honest limits and honest errors: if a file is too big or a format is unsupported, we say so immediately, in plain words.
• Familiar UX patterns. Nothing to learn: if you have seen a file manager and a Download button, you already know how to use nopa.
• Speed is politeness too: pages are light and open fast even on a slow connection.

The tech: the best parts of the modern web

The browser has quietly become a full platform over the past few years, and we squeeze everything we can out of it. WebAssembly and WebGPU run in the browser what used to require a server: nopa removes photo backgrounds and transcribes audio with neural networks right on your device — the model downloads to you, not your files to us. PDFs are merged, split, and compressed right there in the tab.

Results live in the browser’s local storage, and nopa installs as an app (PWA) — some tools work with no internet at all. A nice side effect of this architecture: privacy is not a clause in a policy, it is a physical property of the system. We do not have to promise not to look at your files — we simply do not have them.

The desktop: results should not die

The result of an online tool usually dies the moment you download it: make a QR code — lose the link; compress a PDF — start over tomorrow. That is why nopa has a desktop: everything you make in the tools lands in Files — also in your browser, also yours. From there, a result opens in another app: finish the note, extend the board, shrink the PDF further. You stay in context instead of dragging files between tabs.

What’s next

We will keep improving what is there and carefully extend the lineup — without turning into a landfill of a thousand tools. In the next parts I will go deeper under the hood: why nopa has no backend, how your data lives in the browser, and what running locally means in practice. You can try it right now — here is the full catalog, and ideas and questions are welcome at n1.team.