Skip to content

JoshDoesIT/Sentinellium

Sentinellium

Sentinellium

The Client-Side, Privacy-Preserving AI Defense Grid
Intelligence at the edge. Privacy at the core.

License Version Engines Tests


A WebGPU-accelerated browser extension that runs AI security inference directly on the user's GPU. Zero cloud dependency, zero added latency, complete privacy. Paired with an enterprise management console for fleet-wide threat visibility.

Architecture

sentinellium/
├── apps/
│   ├── extension/     # Chrome/Edge extension (WXT + React 19)
│   ├── console/       # Enterprise dashboard (Next.js 16)
│   └── docs/          # Documentation site
├── packages/
│   ├── engines/       # Shared engine classes (alert, fleet, dashboard)
│   ├── api-client/    # Typed API client for frontend consumption
│   ├── types/         # Shared TypeScript types
│   ├── ui/            # Shared UI components
│   ├── eslint-config/ # Shared ESLint configuration
│   └── typescript-config/ # Shared TS configuration
├── services/
│   └── api/           # Backend API (Hono)
└── docs/              # Reference documents

Core Engines

Engine Purpose Technology
Context Engine Real-time semantic phishing detection ONNX Runtime Web + WebGPU
Provenance Engine C2PA deepfake defense and media verification C2PA JS SDK
DLP Engine Shadow AI data leakage prevention Microsoft Presidio (WASM)

Enterprise Console

The management console provides fleet-wide visibility and enterprise security features:

  • SOC Dashboard — Real-time threat metrics, engine health, trend analysis
  • Alert Investigation — Unified feed with severity filtering, search, and detail views
  • Fleet Management — Monitor all extension instances across the organization
  • Threat Timeline — Temporal analysis with hourly/daily bucketing and trend detection
  • Geographic Intelligence — TLD-based threat origin heatmaps
  • Audit Log — Full compliance trail with actor, action, and IP tracking
  • User Management — RBAC (Admin/Analyst/Viewer), MFA enforcement
  • Authentication — MFA with TOTP, SSO (SAML/OAuth) support
  • Compliance — SOC 2, GDPR, HIPAA framework support

Quick Start

# Install dependencies
pnpm install

# Start the extension in dev mode (auto-opens Chrome)
cd apps/extension && pnpm dev

# Start the enterprise console
cd apps/console && pnpm dev

# Run all tests
pnpm test:run

# Type check
pnpm check-types

# Lint
pnpm lint

Tech Stack

Layer Technology
Extension Framework WXT (Vite-based)
Extension UI React 19 + TypeScript
AI Inference ONNX Runtime Web (WebGPU)
Enterprise Console Next.js 16 (App Router)
API Hono
Database Prisma + PostgreSQL
Styling Vanilla CSS (Brand System)
Testing Vitest + Playwright
Monorepo Turborepo + pnpm

Development

This project follows strict Test-Driven Development (TDD). Every feature begins with a failing test. See CONTRIBUTING.md for the full development workflow.

Branch Protection

All PRs to main require:

  • All CI checks passing (lint, typecheck, test, build)
  • CodeQL security scan passing

License

MIT

About

The Client-Side, Privacy-Preserving AI Defense Grid. WebGPU-accelerated browser extension for local AI security inference.

Resources

License

Code of conduct

Contributing

Security policy

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages