AI-powered litigation prevention OS that converts customer reviews into legal intelligence.
Powered by Amazon Nova 2 Lite on AWS Bedrock + OWL Ontology Reasoning.
"Palantir for Reputation & Legal Risk"
Live Demo: https://onto-review.vercel.app
Hackathon: Amazon Nova AI Hackathon 2026 · Category: Agentic AI
Hashtag: #AmazonNova
When K-Beauty and K-Food brands enter the US market, a single review like "this cream burned my skin" can trigger a multi-million dollar product liability lawsuit. Most SMBs discover this risk only after receiving a legal notice — by then, it's too late.
OntoReview detects legal risks hidden in product reviews before they become lawsuits.
Amazon Reviews (ASIN input)
↓
Risk Classification (Amazon Nova 2 Lite)
↓
OWL Ontology Reasoning (owlready2)
↓
Legal Precedent Matching (Micro-RAG)
↓
Financial Exposure Estimation
↓
Risk Mitigation Playbook
↓
Audit Trail (Duty of Care)
Every analyzed review generates:
| Field | Example |
|---|---|
| Risk Category | Product Liability |
| Severity Score | 8 / 10 |
| Confidence | 0.82 |
| Matched Precedent | Johnson & Johnson Talc Case |
| Estimated Legal Exposure | $2,300,000 |
Amazon Nova 2 Lite (amazon.nova-2-lite-v1:0) serves as the primary reasoning engine through AWS Bedrock. Nova powers every intelligent component:
- Risk Classification — analyzes review text, assigns legal risk categories with severity scores (1–10) and confidence levels
- Legal Exposure Estimation — cross-references detected risks with matched legal precedents to estimate financial damages
- Playbook Generation — generates context-aware mitigation strategies per risk category and severity
- Compliance Analysis — evaluates risks against multi-jurisdiction regulations (US FDA, EU Consumer Protection, KR Fair Trade)
- Agent Response Simulation — powers configurable AI agents with different autonomy levels for automated risk response
The system uses a clean provider abstraction layer. Switching between Nova, OpenAI, and Gemini requires only one environment variable:
LLM_PROVIDER=bedrock # Amazon Nova (hackathon / production)
LLM_PROVIDER=openai # OpenAI GPT-4o-mini (fallback)
LLM_PROVIDER=google # Gemini 2.0 Flash (local development)Implementation: core/utils/openai_client.py
1. Risk Intelligence Dashboard
Real-time dashboard displaying total legal exposure, critical risk count, severity distribution, and risk category breakdown.
2. Risk Response Playbook
AI-generated mitigation strategies tailored to each detected risk. Includes priority actions, timeline, and escalation paths.
3. Trust & Safety Audit
Append-only audit log recording every scan, classification, and risk flag. Exportable as PDF for legal compliance (Duty of Care).
4. Domain Ontology Studio
Custom OWL ontology rule editor. Users can define industry-specific risk classification rules that feed into the ontology engine.
5. Global Compliance Tracker
Multi-jurisdiction regulation checking across US, EU, and KR. Maps detected risks to relevant regulatory frameworks.
6. Agent Communication Setup
Configurable AI agent autonomy levels (1–5) for automated risk response. Includes simulation mode for testing agent behavior.
| Layer | Technology |
|---|---|
| LLM | Amazon Nova 2 Lite via AWS Bedrock |
| Ontology | OWL 2 (owlready2) |
| Backend | Python, FastAPI |
| Frontend | React, Tailwind CSS |
| Database | SQLite, SQLAlchemy |
| Legal RAG | In-memory embedding cache + cosine similarity + TF-IDF fallback |
| Deployment | Vercel (frontend), Render (backend) |
OntoReview/
├── backend/
│ ├── main.py # FastAPI application + lifespan
│ ├── routers/ # API route handlers
│ │ ├── risk.py # Risk analysis endpoints
│ │ ├── agent.py # Agent communication setup
│ │ ├── compliance.py # Global compliance tracker
│ │ ├── studio.py # Domain ontology studio
│ │ ├── audit.py # Trust & safety audit
│ │ ├── discovery.py # Web discovery engine
│ │ └── ...
│ ├── services/ # Business logic
│ │ ├── risk_service.py # Core risk analysis (Nova-powered)
│ │ ├── ontology_engine.py # OWL ontology reasoning
│ │ ├── legal_rag_service.py # Legal precedent matching (Micro-RAG)
│ │ ├── playbook_service.py # Risk response playbook generation
│ │ └── ...
│ ├── data/
│ │ └── legal_cases.json # 30 curated US legal precedents
│ └── database/
│ ├── database.py # SQLite connection
│ └── models.py # SQLAlchemy models (audit trail)
├── core/
│ ├── config.py # Environment & LLM configuration
│ └── utils/
│ └── openai_client.py # Multi-provider LLM abstraction (Nova/OpenAI/Gemini)
├── frontend/
│ └── src/
│ ├── components/ # React UI components
│ │ ├── RiskIntelligence.jsx
│ │ ├── RiskPlaybook.jsx
│ │ ├── AuditTimeline.jsx
│ │ ├── OntologyStudio.jsx
│ │ ├── ComplianceTracker.jsx
│ │ ├── AgentSetup.jsx
│ │ └── ...
│ └── api/client.js # API client
├── infra/ # Deployment configs (Nginx, systemd)
├── tests/ # Unit tests
├── CLAUDE.md # Development guide
└── requirements.txt
- Python 3.11+
- Node.js 20+
- AWS Account with Bedrock access (Amazon Nova 2 Lite enabled in us-east-1)
# Clone
git clone https://github.com/heeoneie/OntoReview.git
cd OntoReview
# Backend
python -m venv venv
source venv/bin/activate # Windows: venv\Scripts\activate
pip install -r requirements.txt
# Frontend
cd frontend
npm install
cd ..Create a .env file in the project root:
# AWS Bedrock (Amazon Nova) — Required
AWS_ACCESS_KEY_ID=your_aws_access_key
AWS_SECRET_ACCESS_KEY=your_aws_secret_key
AWS_DEFAULT_REGION=us-east-1
LLM_PROVIDER=bedrock
# Fallback LLMs (Optional)
OPENAI_API_KEY=your_openai_key
GOOGLE_API_KEY=your_google_key# Backend (terminal 1)
uvicorn backend.main:app --reload
# Frontend (terminal 2)
cd frontend
npm run devThe system supports this 3-minute demo flow:
- Input — Enter an Amazon ASIN or use demo data
- Ingest — System loads product reviews
- Analyze — Amazon Nova classifies legal risks with severity scores
- Match — Micro-RAG finds relevant US legal precedents
- Dashboard — Risk Intelligence displays total legal exposure (e.g., $5.4M)
- Playbook — AI generates risk mitigation strategies
- Audit — Immutable audit trail records all events for compliance
RiskScore = Σ(severity_i × confidence_i)
TotalLegalExposure = Σ(estimated_loss_usd_i)
Where severity_i is the Nova-generated risk severity (1–10), confidence_i is classification confidence (0–1), and estimated_loss_usd_i is derived from matched US legal precedent settlement data.
- Product Liability
- Regulatory Risk
- False Advertising
- Consumer Safety
- Class Action Risk
| Method | Endpoint | Description |
|---|---|---|
| POST | /api/risk/demo |
Run risk analysis on demo reviews |
| POST | /api/risk/ontology |
OWL ontology-based risk classification |
| POST | /api/risk/playbook/generate |
Generate risk mitigation playbook |
| GET | /api/risk/ontology/graph |
Get ontology knowledge graph |
| POST | /api/compliance/check |
Multi-jurisdiction compliance check |
| GET | /api/compliance/regulations |
List available regulations |
| POST | /api/agent/simulate |
Simulate AI agent response |
| GET | /api/audit/events |
Retrieve audit trail |
| POST | /api/discovery/search |
Web discovery engine search |
| GET | /api/kpi/summary |
Dashboard KPI summary |
| GET | /api/health |
Health check |
Frontend — Vercel
Backend — Render
# render.yaml
services:
- type: web
name: ontoreview-api
runtime: python
buildCommand: pip install -r requirements.txt
startCommand: uvicorn backend.main:app --host 0.0.0.0 --port $PORT
envVars:
- key: LLM_PROVIDER
value: bedrock
- key: AWS_ACCESS_KEY_ID
sync: false
- key: AWS_SECRET_ACCESS_KEY
sync: false
- key: AWS_DEFAULT_REGION
value: us-east-1MIT License