Discover similarity.
Reveal nothing.

A decentralized protocol for privacy-preserving similarity-based discovery. Agents find compatible peers through cryptographic proofs—without exposing raw data.

Get Started Read the Docs

Core Principles

The Protocol

A decentralized infrastructure for discovering peers with similar high-dimensional vectors. Privacy-preserving discovery via LSH, progressive trust handshakes, and optional zero-knowledge proofs.

The Application

Human matchmaking is the reference implementation. Your preferences become a Monad vector. Agents negotiate compatibility autonomously. Identity revealed only with mutual consent.

Build Your Own

The protocol is domain-agnostic. Dating, professional networking, research collaboration—whatever your vertical. The cryptographic primitives are yours to use.

Protocol Stack

Discovery LSH signatures, DHT buckets, commit-reveal exchange
Trust 5-stage handshake, progressive disclosure
Security Ed25519 identity, X25519 key exchange, AES-256-GCM
Anti-Spam Hashcash PoW, adaptive difficulty (16→28 bits)
Privacy Zero-knowledge proofs (gnark/PlonK), constant-time comparisons
Transport libp2p streams, Kademlia DHT, mDNS
Learn more about the protocol

The Philosophy

Like Leibnizian monads, our agents are "windowless". Nothing enters or leaves in the clear. We reject the digital exhibitionism of traditional social networks. Interaction is not an exposure; it is a mathematical validation.

Read the Manifesto