Reduce certainty, don't erase every trace
A cover message can make the protected payload less obvious. It does not guarantee that a technical analysis, a device or the other person will not reveal the conversation.
Plausible deniability is not invisibility and does not guarantee safety under duress. An unlocked or compromised device, copies in the transport app, screenshots, backups or other clues can still reveal a conversation. Layergram reduces some traces, but does not monitor the operating system, transport apps, or a physical adversary.
The raw encrypted payload contains no plaintext Layergram markers or unencrypted sender or recipient identifiers. A cover message can carry it inside normal-looking text. Deep links intentionally give up this visual deniability in exchange for convenience, because they identify Layergram.
An optional passphrase can derive a separate identity from the same recovery phrase. Layergram does not save that identity like a normal profile and only keeps its keys in memory while the passphrase is active.
When the passphrase identity is not active, the records belonging to it remain encrypted and unreadable in the context of the base identity. Message IDs, timestamps, and auxiliary status are also stored in encrypted local records.
Layergram authenticates encrypted messages to participants, but does not attach a separate digital signature designed to prove authorship to an external observer. The payload alone is therefore weaker evidence of who wrote it.
Frequently asked questions
No. Encryption protects content from someone who doesn't have the key. Deniability, on the other hand, concerns the possibility that the evidence conclusively demonstrates the existence of certain protected data or the paternity of a person.
No. A Layergram link is recognizable and transport metadata, device artifacts, screenshots, backups or the other participant can provide evidence. Deniability concerns specific cryptographic and local signals, not every surrounding trace.
Its derived keys do not remain in memory, the identity does not appear as a normal local profile, and the associated records remain opaque until the correct passphrase is derived from it again. If you do not reveal the passphrase under pressure, it is mathematically impossible to prove that the indecipherable messages in the device's memory are related to that inactive identity.
Not alone. When an endpoint is unlocked, observed or compromised, evidence external to the encrypted payload can reveal contents or patterns of use.
Start free
Start free. Choose one person. Invite them in a few steps.