ProviderTone

AI-powered style profile generation for healthcare providers. Capture your unique communication voice and documentation style to enable AI systems that sound authentically like you.

Overview

ProviderTone helps healthcare providers create detailed style profiles through two methods:

1. Web Application - Interactive interrogation wizard that guides providers through examples and questions to extract their communication patterns
2. Local CLI Tool - Extract profiles from existing clinical notes and portal messages (PHI stays on your infrastructure)

Project Structure

providertone/
├── app/                    # Next.js web application
├── components/             # React components
├── lib/                    # Shared utilities
├── providertone-local/     # Python CLI for local profile extraction
└── docs/                   # Specifications and plans

Web Application

The Next.js web application provides an interactive wizard for providers to create their style profiles.

Quick Start

# Install dependencies
npm install

# Run development server
npm run dev

# Open http://localhost:3000

Features

• Interactive messaging style interrogation
• Documentation style capture
• Real-time profile preview
• JSON export compatible with AI systems
• Demo playground for testing generated profiles

ProviderTone Local (CLI)

For organizations that need to extract profiles from existing clinical data without uploading PHI.

Quick Start

cd providertone-local

# Create virtual environment
python -m venv .venv
source .venv/bin/activate

# Install
pip install -r requirements.txt
pip install -e .

# Configure API key
cp .env.example .env
# Edit .env with your ANTHROPIC_API_KEY

# Run extraction
providertone extract \
  --messages exports/messages.csv \
  --notes exports/notes/ \
  --provider-id dr-smith \
  --output profiles/dr-smith.json

Features

• Runs 100% locally - PHI never leaves your infrastructure
• Supports CSV, JSON, and text file exports from any EHR
• Uses Claude API (with BAA) or local LLMs for air-gapped environments
• Outputs website-compatible JSON style profiles
• Built-in PHI detection and automatic redaction

See providertone-local/README.md for full documentation.

Style Profile Format

Both tools generate profiles in the same JSON format:

{
  "exportedAt": "2026-01-11T12:00:00.000Z",
  "providerId": "dr-smith",
  "messaging": {
    "generatedProfile": {
      "surfacePatterns": { ... },
      "toneDimensions": { ... },
      "negativeConstraints": { ... },
      "judgmentPatterns": { ... },
      "signatureMoves": [ ... ],
      "voiceSummary": "..."
    }
  },
  "documentation": {
    "generatedProfile": {
      "structuralPatterns": { ... },
      "voiceDimensions": { ... },
      "standardPhrasings": { ... },
      "visitTypeVariations": { ... }
    }
  }
}

Use Cases

• AI Draft Generation - Enable AI to write portal messages in the provider's voice
• Documentation Assistance - Generate clinical notes matching provider style
• Quality Consistency - Maintain voice across AI-assisted communications
• Training Data - Create examples for fine-tuning models

Technology Stack

Web Application

• Next.js 14+ (App Router)
• React 18+
• TypeScript
• Tailwind CSS
• Vercel deployment

Local CLI

• Python 3.9+
• Click (CLI framework)
• Anthropic SDK (Claude API)
• Rich (terminal UI)
• Pandas (data processing)

Development

Web Application

npm run dev       # Development server
npm run build     # Production build
npm run lint      # Run linter

Local CLI

cd providertone-local
source .venv/bin/activate

pip install -e ".[dev]"   # Install with dev dependencies
pytest                     # Run tests
black src/                 # Format code
ruff src/                  # Lint

Documentation

• Local Profile Extraction Spec
• Style Profile Integration Plan
• Demo Playground Feature

License

MIT License