Is Sendant end-to-end encrypted?
Yes. Messages are encrypted on your device before delivery. Sendant servers handle ciphertext, not message text.
Private keys stay on your device. Safety numbers, compared in person or on a call, help people confirm who is on the other end. The full architecture — including what the server can still observe — is on the security page, and the browser's weaker trust model is set out separately in security in the browser.
Which encryption does Sendant use?
X3DH for the initial key agreement and the Double Ratchet for ongoing messages — the same primitives Signal uses.
One Go implementation of both ships everywhere: compiled natively into the Android and iPhone apps, and to WebAssembly in the browser. Using well-understood primitives is a deliberate choice, and so is publishing the protocol and its test vectors so the behaviour can be checked even though the source is private — see how to verify a messenger without its source code.
Can Sendant's servers read my messages?
No. Message content is encrypted on your device before it is sent, so the server stores and forwards ciphertext.
A message waiting in the offline mailbox is an encrypted envelope, not readable text — the mailbox is what makes delivery survive an offline recipient, and it does not require decrypting anything. What the server does see is what any server sees to route a message: that a connection happened, and the network address it came from.
Does Sendant hide my IP address?
No. Sendant does not claim to hide network-level metadata such as your IP address, and it does not use onion routing.
This is the most common thing a privacy messenger is assumed to do and does not. Hiding the network address takes a different architecture with its own costs, and pretending otherwise would be the kind of claim that reads well until someone checks. If an adversary who can watch your traffic is in your threat model, treat Sendant as solving the identity problem, not the network one. The security page says so in the same breath as everything else.
How do I check I am talking to the right person?
Compare safety numbers in person or on a call. Every other signal — a name, an avatar, an invite link — can be forged; that one cannot.
End-to-end encryption protects the channel, not the identity at the other end of it. Safety-number verification closes that gap for the conversation you care most about. It costs a minute, and it is the difference between "encrypted" and "encrypted with the person I think it is".
Has Sendant been independently audited?
Not yet. Sendant is built on X3DH and Double Ratchet primitives, and independent audit work is a launch gate before broad privacy-community promotion.
Until then, Sendant should be evaluated as an early private messenger with transparent limits, not as “battle-tested” or “most secure.” If completed third-party audits are a hard requirement for you today, Signal has years of them and wins that row outright.
Is Sendant open source?
No. The source code is private. The trust story is "verifiable, not open".
What that means concretely: the architecture and protocol are publicly documented, test vectors are published, there is a signed direct Android APK whose certificate fingerprint can be checked against DNS, and behaviour can be observed rather than assumed. What it does not mean is that you can read or compile the code — so please do not treat any claim that you can as accurate. Sendant's own pages avoid the phrase for that reason.
Does Sendant track me?
The apps ship with no analytics or crash-reporting SDK. Privacy-respecting analytics run only on this marketing site, never in the app.
That distinction is deliberate: a messenger that knows who you talk to does not need a third-party SDK in the same process, and shipping one would undo the design. On the site, the beacon is aggregate page-view counting for the pages you are reading now — not in anything you send.