Summary

AIVPN addresses the challenge of traditional VPN protocols like WireGuard and OpenVPN being easily detected by Deep Packet Inspection (DPI) systems, which analyze traffic patterns rather than encrypted content. It provides a solution by disguising encrypted VPN traffic as legitimate application data, rendering it invisible to ISPs and firewalls. This project is targeted at users requiring robust privacy and circumvention capabilities against sophisticated network surveillance.

How It Works

The core innovation is the "Neural Resonance" AI module, which embeds small, trained micro-neural networks (MLPs) directly into the binary. These networks analyze real-time UDP packet entropy and inter-arrival times to detect DPI probing or throttling. Upon detecting a compromised traffic mask (e.g., WebRTC), the system seamlessly rotates to a backup mask (e.g., DNS over UDP) without interrupting the connection. This approach is complemented by a patented routing method (USPTO application No. 19/452,440) for unsupervised multi-task routing via signal reconstruction resonance.

Quick Start & Requirements

Pre-built binaries are available for macOS (aivpn-macos.dmg, ~1.8 MB), Windows (aivpn-client.exe, ~6.4 MB + wintun.dll), and Android (aivpn-client.apk, ~6.5 MB). macOS and Linux clients require root privileges for TUN device access. Windows requires the Wintun driver and execution as Administrator. Server deployment is primarily supported on Linux via Docker or bare metal, requiring openssl, sysctl, and iptables for setup. Building from source requires Rust 1.75+.

Highlighted Details

• AI-Driven Traffic Obfuscation: Utilizes "baked" micro-NNs for real-time analysis and dynamic masking of VPN traffic.
• Stealth Operation: Achieves invisibility to DPI by mimicking common application traffic patterns.
• Connection Stability: Seamless mask rotation ensures zero disconnects even when traffic patterns are flagged.
• Modern Cryptography: Implements Zero-RTT, Perfect Forward Secrecy (PFS), and O(1) cryptographic session tags.
• Performance: Written in Rust, resulting in small (~2.5 MB client binary) and efficient applications suitable for low-resource environments like $5 VPS.

Maintenance & Community

The project appears actively maintained by the sole developer "infosave2007". The README does not detail community channels (e.g., Discord, Slack) or specific contribution guidelines beyond general PR welcome messages.

Licensing & Compatibility

The project is licensed under the permissive MIT License. This license allows for free use, modification, and distribution, including integration into proprietary or commercial software without significant restrictions.

Limitations & Caveats

The Windows client is currently in testing. Server-side functionality is primarily developed for and tested on Linux. Client applications necessitate elevated privileges (root/Administrator) for network interface configuration, which may pose security considerations in certain deployment scenarios.