Governed Intelligence - AI Agent Platform powered by Vorion BASIS
Aurais is an AI agent governance platform implementing the Vorion BASIS (Behavioral Agent Standard for Integrity and Safety) specification. It demonstrates real-time trust scoring, capability gating, and cryptographic audit trails for AI agents.
- Trust Scoring: Dynamic 0-1000 scale with 6 tiers (BASIS specification)
- Capability Gating: Permission control based on trust levels
- Audit Trail: Cryptographic proof chain for all decisions
- Multi-LLM Support: Claude, Gemini, Grok integration
- Real-time Governance: HITL oversight with progressive autonomy
| Level | Name | Score Range | Capabilities |
|---|---|---|---|
| L0 | Sandbox | 0-99 | Isolated testing only |
| L1 | Provisional | 100-299 | Basic read; internal messaging |
| L2 | Standard | 300-499 | Standard ops; limited external |
| L3 | Trusted | 500-699 | Extended ops; external APIs |
| L4 | Certified | 700-899 | Privileged ops; financial |
| L5 | Autonomous | 900-1000 | Full autonomy within policy |
- Node.js 18+
- npm or pnpm
# Clone the repository
git clone https://github.com/BanquetAI/trustbot.git
cd trustbot
# Install dependencies
npm install
cd web && npm install && cd ..
# Start API server
npm run api
# Start web frontend (separate terminal)
cd web && npm run dev- Web App: https://aurais.agentanchorai.com
- Governance Engine: https://cognigate.dev
- Trust Scoring SDK:
npm install @vorionsys/atsf-core
┌─────────────────────────────────────────────────────────────────┐
│ Aurais Console (Web) │
│ ┌──────────┐ ┌──────────┐ ┌──────────┐ ┌──────────┐ │
│ │ Agent │ │ Task │ │ Trust │ │ Audit │ │
│ │ Overview │ │ Pipeline │ │ Metrics │ │ Trail │ │
│ └──────────┘ └──────────┘ └──────────┘ └──────────┘ │
└────────────────────────────┬────────────────────────────────────┘
│ REST/WebSocket
┌────────────────────────────┴────────────────────────────────────┐
│ Aurais API │
│ ┌──────────────────────────────────────────────────────────┐ │
│ │ Governance Pipeline (BASIS) │ │
│ │ ┌──────────┐ ┌──────────┐ ┌──────────┐ │ │
│ │ │ INTENT │→ │ ENFORCE │→ │ PROOF │ │ │
│ │ └──────────┘ └──────────┘ └──────────┘ │ │
│ └──────────────────────────────────────────────────────────┘ │
│ ┌──────────────────────────────────────────────────────────┐ │
│ │ Services: Trust, Tasks, Agents, Governance │ │
│ └──────────────────────────────────────────────────────────┘ │
└─────────────────────────────────────────────────────────────────┘
│
┌────────────────────────────┴────────────────────────────────────┐
│ Cognigate Engine │
│ (BASIS Implementation) │
│ • Trust score calculation (0-1000) │
│ • Capability permission matrix │
│ • Cryptographic proof chain │
│ • Policy enforcement │
└─────────────────────────────────────────────────────────────────┘
Aurais is part of the Vorion AI governance ecosystem:
| Component | Description | URL |
|---|---|---|
| BASIS | Behavioral Agent Standard for Integrity and Safety | vorion.org/basis |
| atsf-core | TypeScript SDK for trust scoring | npm |
| Cognigate | Governance engine API | cognigate.dev |
| Aurais | Reference implementation | aurais.agentanchorai.com |
| AgentAnchor | Enterprise governance platform | agentanchorai.com |
# Run tests
npm run test:run
# Type check
npm run build:strict
# Build for production
npm run build:prod
# Deploy to Vercel
cd web && vercel --prodMIT License - see LICENSE for details.
Built with Vorion BASIS - Learn more at vorion.org