In development. RVNT is pre-release — not yet security-audited. Source code, public builds, and the iOS / App Store release aren’t available yet. See the roadmap →

On-Device Defenses

Proof-of-Work Identity

Also known as: PoW identity, computational identity, Sybil-resistant username, proof-of-work username, mining a username

Proof-of-work identity claims a username by spending measurable computation instead of a phone number, email, or ID. To register, your device must find a hash that meets a difficulty target — cheap to verify, expensive to mass-produce. This makes creating thousands of fake accounts costly, providing Sybil resistance without any central registrar or KYC, and without tying you to a real-world identifier.

In a system with no phone number and no email, what stops one person from grabbing ten thousand usernames and flooding the network with fake accounts? That flood is a Sybil attack, named in John Douceur's 2002 paper The Sybil Attack, which proved that without a central authority or some scarce resource, one entity can always forge many identities. Proof-of-work identity supplies the scarce resource: computation.

The mechanism mirrors the puzzles in cryptocurrency mining. To claim a name, your device repeatedly hashes the username plus a changing nonce until it finds an output below a difficulty target — for example, a hash with a required number of leading zero bits. Finding such a hash takes many guesses (real CPU time and electricity), but anyone can confirm it in a single hash. This asymmetry is the whole point: expensive to produce, trivial to verify. The identity server only stores and checks the resulting proof; it never sees a key, contact, or message.

The tradeoff is honest. Proof-of-work raises the cost of mass registration but does not make it impossible — a well-funded attacker with a GPU farm can still grind out names. It is a speed bump and an economic deterrent, not a hard wall. What it buys you is the ability to have stable, claimable usernames with no KYC, no SIM, and no real-world identifier linking the account to you — a deliberate privacy choice that phone-number-based apps cannot offer.

How it works

Registration is a puzzle search. The client constructs a candidate (the desired username plus account public key plus a nonce), hashes it, and checks whether the digest meets the difficulty target. If not, it increments the nonce and tries again — millions of times if needed. When a qualifying nonce is found, the client submits the username, public key, and nonce to the identity server.

The server's job is cheap: it recomputes the single hash, confirms it meets the target, confirms the username is unclaimed, and stores the public mapping. Difficulty can be tuned upward over time or per-name to keep registration costly as hardware improves. Because the proof is bound to the specific username and key, an attacker cannot reuse one solved puzzle to claim a different name.

How RVNT uses Proof-of-Work Identity

RVNT generates a local Ed25519 identity keypair on-device, then claims a human-readable username by completing a proof-of-work puzzle — no phone number, email, SIM, or KYC. The only server RVNT runs stores the resulting public prekey bundles and username-to-key mappings and serves bootstrap peer discovery over a Kademlia DHT; it never sees private keys, contacts, or message content. See the server and architecture docs.

Frequently asked questions

Why does RVNT use proof-of-work instead of a phone number or email?

Phone numbers and emails tie an account to a real-world, surveillable identifier and require KYC-style verification. Proof-of-work provides Sybil resistance — making mass fake-account creation costly — without collecting any personal data. You prove you spent computation, not that you are a particular person.

Does proof-of-work identity make me anonymous?

It avoids linking your account to a phone number, email, or ID, which removes a major correlation point. But anonymity also depends on network-layer protections like Tor and sealed sender. Proof-of-work addresses the registration layer, not traffic analysis or endpoint compromise.

Can someone still create thousands of fake RVNT accounts?

With enough computing power, yes — proof-of-work is an economic speed bump, not an absolute barrier. It makes large-scale account farming expensive and slow rather than free and instant, which deflects casual abuse while preserving a registrar-free, no-KYC design.

Every definition here describes something RVNT actually ships — a post-quantum, end-to-end-encrypted, peer-to-peer messenger with no phone number and no servers.