moa-telehealth 1.0.0

Magnetic Outlier Agent for Telehealth Compliance

MOA Telehealth - Magnetic Outlier Agent

Consent Risk MVP for Telehealth Compliance

MOA – Magnetic Outlier Agent

Stop letting 0.3% of your chunks hallucinate your LLM.

→ Download MOA for RAG Teams (1-page PDF)
→ pip install moa-telehealth
→ Patent Pending • BUSL 1.1 until Nov 2029

Works with LangChain • LlamaIndex • Haystack • Claude Desktop • Cursor

Overview

MOA uses AI-based anomaly detection to flag telehealth visits lacking required consent elements and RAG chunks that cause hallucinations.

Impact: MOA dropped 22% of bad chunks on a test RAG index without touching the rest.

Key Features

• Stage I (Geometric): Multi-metric outlier detection (centroid distance, local density, clustering coefficient)
• Stage II (Fourier): Sequential degradation detection via spectral analysis
• Interactive Demo: Client-side web interface for real-time compliance checking
• Automated Reporting: PDF compliance reports with visualizations

Quick Start

# Clone the repository
git clone https://github.com/grzywajk-beep/moa_v1.git
cd moa_v1

# Install dependencies
pip install -r requirements.txt

# Run the demo
python src/demo.py

Project Structure

moa_v1/
├── src/
│   ├── core/              # Core MOA algorithms
│   │   ├── geometric.py   # Stage I: Geometric detection
│   │   └── fourier.py     # Stage II: Spectral analysis
│   ├── utils/             # Utilities
│   ├── api/               # FastAPI backend
│   └── demo.py            # Demo script
├── web/
│   ├── index.html         # Interactive landing page
│   └── assets/            # CSS, JS, images
├── docs/
│   ├── presentation/      # Course project materials
│   └── reports/           # Generated compliance reports
├── tests/
└── data/
    └── synthetic/         # Sample datasets

Performance Benchmarks

Metric	MOA (λ=1.5)	LOF	Isolation Forest
Precision	0.92	0.73	0.81
Recall	0.87	0.68	0.75
F1 Score	0.89	0.70	0.78

Roadmap

• Core algorithm implementation
• Synthetic data generation
• Interactive web demo
• EHR integration (Epic, Cerner)
• Real-time compliance alerts
• Mobile app (iOS/Android)

License

Business Source License 1.1 - See LICENSE for details.

Commercial Use:

• Free for evaluation, internal use, and research.
• Commercial License Required for:
  • Offering MOA as a paid hosted service (SaaS).
  • Embedding MOA in a commercial product for external users.
• Converts to Apache 2.0 on 2029-11-24.

See MOA_Enterprise_Licensing_Overview.md for details.

Contact

Michael Ordon - GitHub: grzywajk-beep

---

Transforming telehealth compliance from a tedious afterthought into a streamlined, intelligent process.