Next-gen proxy that adapts to bypass censorship
Sush Core is a censorship circumvention tool built for hostile network environments. It combines quantum-safe crypto with adaptive protocols to stay ahead of detection systems.
- Quantum-Safe Crypto: Uses ML-KEM (Kyber) - your keys won't break when quantum computers arrive
- Protocol Chameleon: Switches between QUIC, WebSocket, TCP, UDP based on what works
- Hidden Channels: Tunnels data through DNS queries, NTP packets, and TTL manipulation
- Smart Detection: ML models learn censorship patterns and adapt automatically
- Onion Routing: Multi-hop encryption like Tor, but harder to block
- Traffic Disguise: Makes your packets look like regular web browsing
git clone https://github.com/soroushdeimi/sush-core.git
cd sush-core
pip install -r requirements.txt
# Start the client proxy
python sush_cli.py client --config config/client.conf
# Point your browser to localhost:8080 (SOCKS5)
# Copy and edit server config
cp config/server.conf.example config/server.conf
vim config/server.conf # Add your server details
# Start serving
python sush_cli.py server --config config/server.conf
export SUSH_SERVER_HOST=your-server.com
export SUSH_SERVER_PORT=9090
export SUSH_THREAT_LEVEL=high # low, medium, high, paranoid
export SUSH_ML_ENABLE=true
config/client.conf
- Client settingsconfig/server.conf
- Server settings
Check USER_GUIDE.md for detailed configuration options.
Sush Core has four main layers:
- Core: Quantum crypto, adaptive encryption, traffic morphing
- Transport: Protocol switching, steganographic hiding
- Network: Multi-hop routing, node verification
- Control: ML adaptation, censorship detection
Read ARCHITECTURE.md for the technical deep-dive.
# Run all tests
python run_tests.py
# Quick smoke test
python tests/test_core_components.py
Found a bug? Want to add a feature? Here's how:
- File issues on GitHub Issues
- Submit pull requests for features
- Join discussions in GitHub Discussions
- Bridge discovery protocol
- Mobile apps (iOS/Android)
- Browser extension (WebAssembly)
- Better ML models
- Performance improvements
MIT License - see LICENSE for details.