OMNI

Dual Post-Quantum Secure WireGuard Mesh Networking

ML-KEM + SLH-DSA encryption. Quantum-proof. Zero-config deployment. All architectures.

Core Features

🔐 Dual Post-Quantum Encryption

ML-KEM (key encapsulation) + SLH-DSA (signatures). Both quantum-resistant. Hybrid classical Curve25519 + PQ = forward-secret against quantum computers.

⚡ Zero-Config Enrollment

Ephemeral tokens (one-time use, 32-byte secret, TTL-based). Workers enroll in seconds without manual secrets.

🌐 Multi-Architecture

Linux (x64/ARM64/ARM32), Windows, Android, macOS. One binary fits everywhere.

📡 WireGuard Foundation

Kernel-mode VPN. Fast, modern, battle-tested. 1420 MTU, configurable per-node.

🎯 Central Registry + Almanac

Authoritative hive topology. Heartbeat tracking. Role-based peer filtering (coordinator/worker/relay/gateway).

🔗 ComputeEngine Integration

Exports peers.json for automatic worker binding. HMAI coordinator manages hive enrollment.

Specifications

Encryption ML-KEM + SLH-DSA (Dual PQ)

License MIT (Open Source)

Language Rust 1.75+

Binary Size 6.8MB (Linux x64)

Dependencies 40 crates (rustls, tokio, warp)

Build Time ~2 minutes (native)

Network WireGuard (51820/UDP)

Registry API HTTP REST (:9999)

Version v0.1.0 (2026-09-12)

Supported Platforms & Architectures

✅ Linux x86_64

✅ Windows x64

🔧 Linux ARM64

🔧 Linux ARM32

📦 Android ARM64

📦 Android ARM32

🖥️ macOS x64

🖥️ macOS ARM64

✅ Built & Tested | 🔧 Target Ready | 📦 NDK Required | 🖥️ Native Compile

Quick Start

# Coordinator (Newton CT115)
wget https://github.com/mradamantware/OMNI/releases/download/v0.1.0/omni-linux-x64
chmod +x omni-linux-x64
sudo ./omni-linux-x64 registry --listen 0.0.0.0:9999

# Issue enrollment token
./omni-linux-x64 issue-token --name worker-1 --ttl 3600

# Worker enrolls & joins mesh
./omni-linux-x64 enroll-token --token-id UUID --token-secret SECRET
sudo ./omni-linux-x64 start --config ~/.omni/config.toml
        

Use Cases

🧠 HMAI Hive

Distributed AI training. Newton brain coordinates workers over OMNI mesh. Automatic peer discovery.

🎮 Demerzel GPU

Multi-node GPU inference. Workers coordinate tensor distribution via mesh. No external network needed.

🖥️ ComputeEngine

Bind distributed workers automatically. CE reads peers.json, configures mesh, scales on demand.

📱 Edge / IoT

Deploy on Raspberry Pi, ARM boards, mobile devices. Same binary, instant mesh formation.

🏢 Enterprise

Windows + Linux infrastructure. Secure cross-platform coordination. Post-quantum ready.

🌍 Distributed Systems

Any topology. Any architecture. Mesh scales to thousands of nodes. Quantum-resistant.

Downloads & documentation on NoHatHacker →Talk to us