Trust & Safety
Security at DiemDesk
We built DiemDesk around a single idea: your files are yours. Most tools run entirely inside your browser, so your documents never reach our servers in the first place. Here is exactly how everything works — in plain language.
Last updated June 2026
Our promise
Your files never leave your device unless you choose — and even then, only you hold the key. We design every feature so that we see as little of your data as technically possible. The best way to keep something private is to never receive it.
How our tools work
Tools like Merge, Split, Compress, Convert and QR generation run 100% inside your browser, using your own device’s processing power. When you drop a file in, it is opened locally — it is never uploaded to us. When the tool finishes, the result downloads straight back to you.
Zero uploads for in-browser tools
There is no copy of your file on our servers — not during processing, not after. Close the tab and there is nothing left anywhere but on your own computer. No accounts required, no watermarks, and no daily limits — just generous file sizes (competitors cap you at ~15 MB or a couple of tasks a day).
Where your data goes
Different tools handle your files differently — and we’re upfront about every one. Here is exactly where your data goes, by tool type:
The technology we use
- In-browser processing. File operations run on your device using modern web technology (WebAssembly and the browser’s native APIs) — the same engine that powers the page you’re reading.
- AES-256 encryption. The File Vault uses AES-256-GCM — the same encryption standard trusted by banks and governments — applied on your device, in your browser, via the standard Web Crypto API.
- Strong key derivation. Your vault key is derived from your passphrase using Argon2, a modern, deliberately slow function that makes guessing attacks impractical.
- Encrypted connections. All traffic to DiemDesk is protected with TLS (HTTPS), and our servers are hardened and kept up to date.
Built to run light on your device
Because your files are processed on your device — never uploaded — the work runs on your own CPU and memory. We engineered everything to keep that light:
- Handled in chunks. Files are processed in small pieces, and memory is released as soon as each part is done — nothing piles up.
- Fast and responsive. The heavy lifting runs on WebAssembly, and the app pauses between steps so your browser stays smooth while it works.
- Sensible limits. Caps on file size and rendering quality stop any single job from running away with your device’s resources.
One thing about large files
Big files use more of your own device’s power — that’s the price of keeping them off our servers and fully private. For everyday documents, you won’t notice a thing.
Working offline
Not yet — and we’d rather say so than fudge it. Your files are processed on your device, and that part is true today; you can watch it in your browser’s network tab. But the app itself is still loaded from the web each time you open it, so if you’re offline you get a blank page.
Why it isn’t on right now
We shipped an offline cache in July 2026 and pulled it a few days later: it was serving some people an outdated copy of the site. We’d rather ship nothing than ship something stale. It’s being rebuilt properly, and this page will change the day it lands.
When it does return, it will be per-tool and per-browser: once you’ve actually used a tool online, your browser keeps the pieces that tool needed, and it can run without a connection after that. A tool you’ve never opened won’t work offline, because its files were never fetched.
Worth knowing either way: Office conversions, OCR, PDF/A, Link in bio, the AI tools (Summarize, Chat with your PDF, Translate) and File Vault sync run on our servers and will always need a connection. Every tool in the catalog is labelled.
How we handle AI features
AI tools — Summarize, Chat with your PDF, and Translate — do need to read the text of your document to work, so they can’t run fully offline. Here is exactly what happens, and what doesn’t:
- 1
You choose to use AI
AI never runs automatically. It only happens when you click Summarize, ask a question, or request a translation.
- 2
Your document’s text is sent for that one request
The relevant text is sent over an encrypted connection to our AI provider, processed in real time, and the answer is returned to you.
- 3
It is not stored or used to train models
We do not keep a copy of your document to power AI, and your content is not used to train AI models.
Where AI is different
Unlike our in-browser tools, AI features do transmit your text for the moment it takes to answer — that’s simply how AI works. The difference with DiemDesk is that it’s always your choice, it’s used only for your request, and it’s never retained or sold.
File Vault — end-to-end encrypted storage
Most of DiemDesk never stores your files. The File Vault is the one exception — and it’s built so that even we can never read what’s inside it. This is called zero-knowledge (or end-to-end) encryption, the same approach used by privacy-first services like Proton Drive and Tresorit.
- 1
Encrypted on your device, before upload
When you save a file to your Vault, your browser scrambles it with AES-256 first. What gets uploaded is unreadable gibberish.
- 2
We store only the gibberish
Our servers hold the encrypted version. We have no way to unscramble it — even if we were asked to, or our servers were breached, there is nothing readable to hand over.
- 3
Only you can unlock it
When you open your Vault on any device, your browser rebuilds your key from your passphrase and decrypts your files locally. The key never reaches us.
What this means for you
- We cannot read, scan, or share your vaulted files — it’s technically impossible for us.
- A data breach of our servers would expose only encrypted gibberish, useless without your key.
- Your vault passphrase is separate from your login password. Logging in proves who you are; your vault passphrase unlocks your files and is never sent to us.
- One passphrase unlocks your whole vault — you don’t need a separate password per file.
Important: if you lose your passphrase, your files cannot be recovered
Please read this carefully
Because your File Vault is encrypted so that only you can open it, we never receive or store your vault passphrase. That is what makes it truly private — but it also means:
If you forget your vault passphrase, we cannot reset it, and your vaulted files cannot be recovered by anyone — including us. They are permanently lost.
This works just like a cryptocurrency wallet or a true zero-knowledge service: total privacy means total responsibility for your passphrase.
Your safety net: the recovery key
When you first set up your Vault, we give you a one-time recovery key — a long code you should save somewhere safe (a password manager, or printed and stored offline). If you ever forget your passphrase, the recovery key is the only way back into your Vault. Keep it safe and never share it.
Our infrastructure
Beyond in-browser privacy, we keep our own systems tight: encrypted connections (TLS) everywhere, hardened and regularly patched servers, restricted administrative access, and monitoring for unusual activity. Account passwords are stored only as salted hashes — never in plain text.
Found a security issue?
We welcome responsible disclosure. If you believe you’ve found a vulnerability, please email [email protected] with the details. We’ll investigate promptly and keep you updated.