TrustEdge — Trustable Edge AI (Rust)

TrustEdge is a professional-grade Rust platform for trustable edge AI with privacy-preserving, provenance-aware data processing. Now hosted at TrustEdge-Labs, the project has evolved from a personal learning journey into a comprehensive system for secure edge computing.

🏢 Professional Organization

TrustEdge-Labs GitHub organization reflects the project's evolution to professional-grade development with dedicated governance, community management, and long-term sustainability.

✨ Key Features

🎵 Live Audio Capture

  • Real-time microphone input with cross-platform support (Linux/ALSA, Windows/WASAPI, macOS/CoreAudio)
  • Configurable quality with device selection, sample rates, and channel configuration
  • Immediate encryption of audio chunks before leaving the device

🔧 Universal Backend System

  • Capability-based crypto operations supporting multiple backends
  • ✅ Keyring Backend (OS keyring integration) - Available now
  • 🔄 YubiKey/TPM/HSM Backends - Planned with ready architecture
  • Runtime discovery and preference-based selection

🔐 Enhanced Security

  • Ed25519 mutual authentication between clients and servers
  • Session management with cryptographic session IDs and timeouts
  • AES-256-GCM encryption with signed manifests for provenance
  • Connection resilience with retry logic and graceful shutdown

📋 Format-Aware Processing

  • 30+ file type support with MIME detection
  • Format preservation during encryption/decryption cycles
  • Audio metadata preservation (sample rate, channels, bit depth)
  • Inspection tools to view format without decryption

🚀 Architecture Overview

Data Sources → Format Detection → Encryption Pipeline → .trst Archives
     ↓              ↓                    ↓                    ↓
  Files         MIME Types         AES-256-GCM         Network Transport
Live Audio    Audio Metadata     Ed25519 Sigs        Mutual Auth
 Sensors      Type Preservation   Manifest Binding    Session Mgmt

🧪 Quality Assurance

93 comprehensive tests with 100% success rate covering: - 53 Unit Tests: Core functionality + Universal Backend + Software HSM - 15 Roundtrip Tests: End-to-end encryption/decryption workflows - 9 Software HSM Tests: Cross-session persistence and CLI workflows - 7 Network Tests: Real client-server communication - 6 Universal Backend Tests: Capability-based workflows - 3 Authentication Tests: Mutual authentication and session management

🔧 Quick Start

Installation

# Basic installation (no audio)
git clone https://github.com/TrustEdge-Labs/trustedge.git
cd trustedge/trustedge-audio
cargo build --release --no-default-features

# Full installation (with live audio)
cargo build --release --features audio

Live Audio Capture

# Capture 10 seconds of live audio
./target/release/trustedge-audio \
  --live-capture \
  --envelope voice_note.trst \
  --key-out voice_key.hex \
  --max-duration 10

# List available audio devices
./target/release/trustedge-audio --list-audio-devices

Network Mode with Authentication

# Start authenticated server
./target/release/trustedge-server \
  --listen 127.0.0.1:8080 \
  --require-auth \
  --decrypt

# Connect authenticated client
./target/release/trustedge-client \
  --server 127.0.0.1:8080 \
  --input file.txt \
  --require-auth

📚 Professional Documentation

🎯 Current Status

✅ Phase 1: Foundation (Core encryption/decryption) - COMPLETED
✅ Phase 2: Key Management (Universal Backend system) - COMPLETED
✅ Phase 3: Network Operations (Mutual authentication) - COMPLETED
🔄 Phase 4: Live Audio Capture (Real-time processing) - IN PROGRESS
📋 Phase 5: Hardware Security (YubiKey, TPM, HSM) - PLANNED

🤝 Contributing

TrustEdge welcomes contributions! See the comprehensive guidelines: - Contributing Guide - Development Setup - Project Board for current tasks

📄 License

Mozilla Public License 2.0 (MPL-2.0)

From personal learning project to professional trustable edge AI platform - privacy and trust at the edge, powered by Rust.