README.md

---
title: SentinelRx ADR Intelligence Prototype
emoji: 🧠
colorFrom: indigo
colorTo: purple
sdk: gradio
sdk_version: 6.14.0
app_file: app.py
pinned: false
---

# SentinelRx ADR Intelligence Prototype

This Hugging Face Space demonstrates the AI workflow layer for SentinelRx, a patient-facing adverse drug reaction detection companion.

## What the prototype does

The Space lets a user select or paste a patient ADR testimony. It then shows:

1. Severity rating output
2. Medical information extraction output
3. Highlighted testimony evidence
4. Agentic triage recommendation
5. Structured ADR report preview

## Important note

The severity model and medical information extraction model are currently placeholders.

Replace these functions in `app.py` when the final models are ready:

```python
run_severity_model(text)
run_medical_extraction_model(text)
```

The rest of the prototype can stay the same.

## Demo scenarios included

1. Severe reaction from patient testimony
2. Moderate ADR with uncertainty and missing information
3. Hidden safety risk in plain-language testimony

## How to run locally

```bash
pip install -r requirements.txt
python app.py
```

## Hugging Face Space setup

1. Create a new Hugging Face Space
2. Choose Gradio as the SDK
3. Upload:
   - `app.py`
   - `requirements.txt`
   - `README.md`
4. The Space should build automatically

## Presentation framing

SentinelRx is a patient-facing ADR companion. The Hugging Face Space demonstrates the intelligence layer that would power the product: patient testimony goes in, severity and medical entities are extracted, then an agentic triage layer prioritizes the case and prepares an ADR report for human review.