Home / Company / Security and transparency
Security and transparency

CUBRA security: what we see and what we do not

An honest account of what the server sees and what it does not; what is protected and what depends on you.

What we see

Minimal service data

  • Email address for sign-in and receipts
  • Public device keys – so that people can write to you
  • Encrypted messages in transit while the recipient is offline
  • An address hash, kept no longer than 48 hours – against mass sign-ups
What we do not see

And cannot see

  • The content of messages, files and calls
  • Who exactly is writing to you – the sender is hidden
  • File names, types and sizes
  • Your contacts and your address book
Honestly

Limits we state plainly

The browser is weaker than the app

The code in the browser is sent by the server. For full protection, install the app for your computer or phone.

Infected device

If a device has a virus, no app can provide full protection. CUBRA warns about the risks and does not issue keys to an unsafe device.

Independent audit – still ahead

The core has been checked with thousands of tests and mathematical proof models. We will have it checked by independent experts before release and publish the result.

Threat model

What we protect against

  • Against someone intercepting network traffic: in transit there is only ciphertext.
  • Against someone who has seized or hacked a server: there are no keys on the server, and the sender is hidden.
  • Against an administrator or contractor who wants to read messages: access to management does not give access to content.
  • Against someone who has copied a key: keys are renewed, and within a few messages the copy becomes useless.
  • Against someone recording traffic today to read it later with a quantum computer: the key exchange is post-quantum.
  • Against someone trying to forge or replay a message: the app checks authenticity and rejects replays.
  • Against mass attacks on availability: our own DDoS protection in the CUBRA Node.

What the cipher will not protect against

We say this plainly so that you can close these risks by other means. The cipher will not stop the other person from photographing the screen with another phone or forwarding the text. An infected device sees everything its owner sees. An unlocked phone in someone else’s hands reveals whatever is already open on it. That is why the app warns about device risks, and a lost device can be revoked.

How we test our protection

  • Thousands of automated tests on every change, including attack tests.
  • Checks for timing leaks.
  • Mathematical proof models of the key exchange protocol.
  • Internal cryptanalysis of our cipher and a blind attack team: the cipher was attacked with every known method, including quantum ones – no way to break it was found.
  • Independent audit before release; we will publish the findings.

Signed apps and updates

Every CUBRA app and every update is signed with the release key. The app installs only verified updates, and you can check files from the downloads page yourself – instructions are on the “Downloads” page. The release key is kept separately from the servers, so seizing a server does not make it possible to slip you a fake app.

Your data after your subscription ends

Messages you have received always open – we never limit decryption. For organisations, the licence works without a network connection and does not depend on our availability: even if the connection to us is lost, your system keeps working.

Frequently asked questions about security

Can you read my messages?

No. The keys exist only on the devices of the people in the conversation. The server stores only ciphertext and does not know who is writing to whom.

Why is your own cipher not a risk?

It does not replace the world standard; it works underneath it as a second independent layer. To read the correspondence, an attacker would need to break both ciphers and all key exchange mechanisms, including the post-quantum ones, at the same time.

Why is there no independent audit yet?

The product has not been released yet. We will carry out the audit before release and publish the findings; until then, we rely on the internal checks described above.

How do I report a vulnerability?

Write to our security address (it will be given below). We will reply within 72 hours and, with your consent, thank you publicly once the issue is fixed.

Requests from authorities

We have no keys to your messages, so we cannot hand over their content – neither on request nor under pressure. In response to a lawful request, we provide only what we have (account creation date, last sign-in). We will publish a transparency report regularly. Learn more: our legal position.

Found a vulnerability?

Write to security@… (the address will be announced) – we will reply within 72 hours, fix the issue and, with your consent, thank you publicly. Please do not disclose details until it is fixed.