AuthScript

AI-powered prior authorization automation that integrates directly into Epic's clinical workflow via CDS Hooks. AuthScript automatically aggregates clinical data, reasons against payer policies, and generates completed PA forms.

Documentation

Quick Links	Description
Architecture Overview	System patterns and data flow
Design Document	Detailed technical design

Quick Start

Prerequisites

Node.js 20+ with npm
.NET 10 SDK
Docker Desktop or Podman Desktop (required)

Note: Docker Desktop (or Podman) must be running before starting the app. The Intelligence and Dashboard services run in containers.

Setup

# Run the setup script (configures secrets automatically)
./scripts/setup.sh

The script configures LLM provider settings via dotnet user-secrets:

Setting	Description	Default
`llm-provider`	LLM backend: `github`, `azure`, `gemini`, or `openai`	`github`
`github-token`	GitHub PAT for GitHub Models (auto-detected from `gh auth`)	—
`azure-openai-key`	Azure OpenAI API key	—
`azure-openai-endpoint`	Azure OpenAI endpoint URL	—
`google-api-key`	Google Gemini API key	—
`openai-api-key`	OpenAI API key	—
`openai-org-id`	OpenAI organization ID (optional)	—

To use a different provider:

# Azure OpenAI
LLM_PROVIDER=azure AZURE_OPENAI_API_KEY=... AZURE_OPENAI_ENDPOINT=... ./scripts/setup.sh
# Google Gemini
LLM_PROVIDER=gemini GOOGLE_API_KEY=... ./scripts/setup.sh
# OpenAI
LLM_PROVIDER=openai OPENAI_API_KEY=... OPENAI_ORG_ID=... ./scripts/setup.sh

IDE Setup

All IDEs use .NET Aspire to orchestrate services. The AppHost project starts all services (Gateway, Intelligence, Dashboard) with proper dependencies.

VS Code (Recommended)

Install the C# Dev Kit extension
Open the prior-auth folder
Open the Command Palette (Ctrl+Shift+P / Cmd+Shift+P)
Run: ".NET: Open Solution" → select orchestration/AuthScript.sln
In the Solution Explorer, right-click AuthScript.AppHost → "Debug" → "Start New Instance"

Alternatively, from the terminal:

cd orchestration/AuthScript.AppHost
dotnet run

The Aspire Dashboard opens automatically at https://localhost:15888 showing all services.

Visual Studio / Rider

Open orchestration/AuthScript.sln
Set AuthScript.AppHost as the startup project
Press F5 (or Run) to start with debugging
The Aspire Dashboard opens automatically with all services

Schema Synchronization

Regenerate types after modifying API contracts:

npm run sync:schemas

This generates:

TypeScript — React Query hooks, Zod schemas (shared/types/, shared/validation/)
Python — Pydantic models from Gateway spec (apps/intelligence/src/models/generated/)
C# — Records from Intelligence spec (apps/gateway/Gateway.API/Models/Generated/)

CI runs schema sync and fails if generated files drift from committed versions.

Gateway OpenAPI Extraction

Gateway uses runtime OpenAPI generation. To extract locally:

# Start Gateway, then fetch spec
curl http://localhost:5000/openapi/v1.json > apps/gateway/openapi.json
npm run sync:schemas

See shared/schemas/README.md for contract ownership.

Services

Service	Technology	Purpose
Gateway	.NET 10, iText7, Polly	CDS Hooks, FHIR aggregation, PDF generation
Intelligence	Python 3.11, FastAPI	Clinical reasoning (GitHub/Azure/Gemini/OpenAI)
Dashboard	React 19, Vite, TanStack	Shadow dashboard + SMART fallback

Testing

# Run all tests
npm run test# Run specific service tests
npm run test:dashboard
dotnet test apps/gateway/Gateway.API.Tests
# Intelligence tests (run inside container or with local Python)cd apps/intelligence && uv run pytest

Environment Variables

For local development, run ./scripts/setup.sh (see Setup above).

For production or CI/CD, set these environment variables:

Variable	Service	Description
`LLM_PROVIDER`	Intelligence	Provider: `github`, `azure`, `gemini`, `openai`
`GITHUB_TOKEN`	Intelligence	GitHub PAT for GitHub Models
`AZURE_OPENAI_API_KEY`	Intelligence	Azure OpenAI API key
`AZURE_OPENAI_ENDPOINT`	Intelligence	Azure OpenAI endpoint URL
`GOOGLE_API_KEY`	Intelligence	Google Gemini API key
`OPENAI_API_KEY`	Intelligence	OpenAI API key
`OPENAI_ORG_ID`	Intelligence	OpenAI organization ID (optional)
`Epic__ClientId`	Gateway	Epic Launchpad client ID
`Epic__FhirBaseUrl`	Gateway	FHIR R4 endpoint (has default)

License

This is a class project for CSE 589 at the University of Washington.

AuthScript

AI-powered prior authorization automation that integrates directly into Epic's clinical workflow via CDS Hooks. AuthScript automatically aggregates clinical data, reasons against payer policies, and generates completed PA forms.

Documentation

Quick Links	Description
Architecture Overview	System patterns and data flow
Design Document	Detailed technical design

Quick Start

Prerequisites

Node.js 20+ with npm
.NET 10 SDK
Docker Desktop or Podman Desktop (required)

Note: Docker Desktop (or Podman) must be running before starting the app. The Intelligence and Dashboard services run in containers.

Setup

# Run the setup script (configures secrets automatically)
./scripts/setup.sh

The script configures LLM provider settings via dotnet user-secrets:

Setting	Description	Default
`llm-provider`	LLM backend: `github`, `azure`, `gemini`, or `openai`	`github`
`github-token`	GitHub PAT for GitHub Models (auto-detected from `gh auth`)	—
`azure-openai-key`	Azure OpenAI API key	—
`azure-openai-endpoint`	Azure OpenAI endpoint URL	—
`google-api-key`	Google Gemini API key	—
`openai-api-key`	OpenAI API key	—
`openai-org-id`	OpenAI organization ID (optional)	—

To use a different provider:

# Azure OpenAI
LLM_PROVIDER=azure AZURE_OPENAI_API_KEY=... AZURE_OPENAI_ENDPOINT=... ./scripts/setup.sh
# Google Gemini
LLM_PROVIDER=gemini GOOGLE_API_KEY=... ./scripts/setup.sh
# OpenAI
LLM_PROVIDER=openai OPENAI_API_KEY=... OPENAI_ORG_ID=... ./scripts/setup.sh

IDE Setup

All IDEs use .NET Aspire to orchestrate services. The AppHost project starts all services (Gateway, Intelligence, Dashboard) with proper dependencies.

VS Code (Recommended)

Install the C# Dev Kit extension
Open the prior-auth folder
Open the Command Palette (Ctrl+Shift+P / Cmd+Shift+P)
Run: ".NET: Open Solution" → select orchestration/AuthScript.sln
In the Solution Explorer, right-click AuthScript.AppHost → "Debug" → "Start New Instance"

Alternatively, from the terminal:

cd orchestration/AuthScript.AppHost
dotnet run

The Aspire Dashboard opens automatically at https://localhost:15888 showing all services.

Visual Studio / Rider

Open orchestration/AuthScript.sln
Set AuthScript.AppHost as the startup project
Press F5 (or Run) to start with debugging
The Aspire Dashboard opens automatically with all services

Schema Synchronization

Regenerate types after modifying API contracts:

npm run sync:schemas

This generates:

TypeScript — React Query hooks, Zod schemas (shared/types/, shared/validation/)
Python — Pydantic models from Gateway spec (apps/intelligence/src/models/generated/)
C# — Records from Intelligence spec (apps/gateway/Gateway.API/Models/Generated/)

CI runs schema sync and fails if generated files drift from committed versions.

Gateway OpenAPI Extraction

Gateway uses runtime OpenAPI generation. To extract locally:

# Start Gateway, then fetch spec
curl http://localhost:5000/openapi/v1.json > apps/gateway/openapi.json
npm run sync:schemas

See shared/schemas/README.md for contract ownership.

Services

Service	Technology	Purpose
Gateway	.NET 10, iText7, Polly	CDS Hooks, FHIR aggregation, PDF generation
Intelligence	Python 3.11, FastAPI	Clinical reasoning (GitHub/Azure/Gemini/OpenAI)
Dashboard	React 19, Vite, TanStack	Shadow dashboard + SMART fallback

Testing

# Run all tests
npm run test# Run specific service tests
npm run test:dashboard
dotnet test apps/gateway/Gateway.API.Tests
# Intelligence tests (run inside container or with local Python)cd apps/intelligence && uv run pytest

Environment Variables

For local development, run ./scripts/setup.sh (see Setup above).

For production or CI/CD, set these environment variables:

Variable	Service	Description
`LLM_PROVIDER`	Intelligence	Provider: `github`, `azure`, `gemini`, `openai`
`GITHUB_TOKEN`	Intelligence	GitHub PAT for GitHub Models
`AZURE_OPENAI_API_KEY`	Intelligence	Azure OpenAI API key
`AZURE_OPENAI_ENDPOINT`	Intelligence	Azure OpenAI endpoint URL
`GOOGLE_API_KEY`	Intelligence	Google Gemini API key
`OPENAI_API_KEY`	Intelligence	OpenAI API key
`OPENAI_ORG_ID`	Intelligence	OpenAI organization ID (optional)
`Epic__ClientId`	Gateway	Epic Launchpad client ID
`Epic__FhirBaseUrl`	Gateway	FHIR R4 endpoint (has default)

License

This is a class project for CSE 589 at the University of Washington.

AuthScript

AI-powered prior authorization automation that integrates directly into Epic's clinical workflow via CDS Hooks. AuthScript automatically aggregates clinical data, reasons against payer policies, and generates completed PA forms.

Documentation

Quick Links	Description
Architecture Overview	System patterns and data flow
Design Document	Detailed technical design

Quick Start

Prerequisites

Node.js 20+ with npm
.NET 10 SDK
Docker Desktop or Podman Desktop (required)

Note: Docker Desktop (or Podman) must be running before starting the app. The Intelligence and Dashboard services run in containers.

Setup

# Run the setup script (configures secrets automatically)
./scripts/setup.sh

The script configures LLM provider settings via dotnet user-secrets:

Setting	Description	Default
`llm-provider`	LLM backend: `github`, `azure`, `gemini`, or `openai`	`github`
`github-token`	GitHub PAT for GitHub Models (auto-detected from `gh auth`)	—
`azure-openai-key`	Azure OpenAI API key	—
`azure-openai-endpoint`	Azure OpenAI endpoint URL	—
`google-api-key`	Google Gemini API key	—
`openai-api-key`	OpenAI API key	—
`openai-org-id`	OpenAI organization ID (optional)	—

To use a different provider:

# Azure OpenAI
LLM_PROVIDER=azure AZURE_OPENAI_API_KEY=... AZURE_OPENAI_ENDPOINT=... ./scripts/setup.sh
# Google Gemini
LLM_PROVIDER=gemini GOOGLE_API_KEY=... ./scripts/setup.sh
# OpenAI
LLM_PROVIDER=openai OPENAI_API_KEY=... OPENAI_ORG_ID=... ./scripts/setup.sh

IDE Setup

All IDEs use .NET Aspire to orchestrate services. The AppHost project starts all services (Gateway, Intelligence, Dashboard) with proper dependencies.

VS Code (Recommended)

Install the C# Dev Kit extension
Open the prior-auth folder
Open the Command Palette (Ctrl+Shift+P / Cmd+Shift+P)
Run: ".NET: Open Solution" → select orchestration/AuthScript.sln
In the Solution Explorer, right-click AuthScript.AppHost → "Debug" → "Start New Instance"

Alternatively, from the terminal:

cd orchestration/AuthScript.AppHost
dotnet run

The Aspire Dashboard opens automatically at https://localhost:15888 showing all services.

Visual Studio / Rider

Open orchestration/AuthScript.sln
Set AuthScript.AppHost as the startup project
Press F5 (or Run) to start with debugging
The Aspire Dashboard opens automatically with all services

Schema Synchronization

Regenerate types after modifying API contracts:

npm run sync:schemas

This generates:

TypeScript — React Query hooks, Zod schemas (shared/types/, shared/validation/)
Python — Pydantic models from Gateway spec (apps/intelligence/src/models/generated/)
C# — Records from Intelligence spec (apps/gateway/Gateway.API/Models/Generated/)

CI runs schema sync and fails if generated files drift from committed versions.

Gateway OpenAPI Extraction

Gateway uses runtime OpenAPI generation. To extract locally:

# Start Gateway, then fetch spec
curl http://localhost:5000/openapi/v1.json > apps/gateway/openapi.json
npm run sync:schemas

See shared/schemas/README.md for contract ownership.

Services

Service	Technology	Purpose
Gateway	.NET 10, iText7, Polly	CDS Hooks, FHIR aggregation, PDF generation
Intelligence	Python 3.11, FastAPI	Clinical reasoning (GitHub/Azure/Gemini/OpenAI)
Dashboard	React 19, Vite, TanStack	Shadow dashboard + SMART fallback

Testing

# Run all tests
npm run test# Run specific service tests
npm run test:dashboard
dotnet test apps/gateway/Gateway.API.Tests
# Intelligence tests (run inside container or with local Python)cd apps/intelligence && uv run pytest

Environment Variables

For local development, run ./scripts/setup.sh (see Setup above).

For production or CI/CD, set these environment variables:

Variable	Service	Description
`LLM_PROVIDER`	Intelligence	Provider: `github`, `azure`, `gemini`, `openai`
`GITHUB_TOKEN`	Intelligence	GitHub PAT for GitHub Models
`AZURE_OPENAI_API_KEY`	Intelligence	Azure OpenAI API key
`AZURE_OPENAI_ENDPOINT`	Intelligence	Azure OpenAI endpoint URL
`GOOGLE_API_KEY`	Intelligence	Google Gemini API key
`OPENAI_API_KEY`	Intelligence	OpenAI API key
`OPENAI_ORG_ID`	Intelligence	OpenAI organization ID (optional)
`Epic__ClientId`	Gateway	Epic Launchpad client ID
`Epic__FhirBaseUrl`	Gateway	FHIR R4 endpoint (has default)

License

This is a class project for CSE 589 at the University of Washington.

AuthScript

AI-powered prior authorization automation that integrates directly into Epic's clinical workflow via CDS Hooks. AuthScript automatically aggregates clinical data, reasons against payer policies, and generates completed PA forms.

Documentation

Quick Links	Description
Architecture Overview	System patterns and data flow
Design Document	Detailed technical design

Quick Start

Prerequisites

Node.js 20+ with npm
.NET 10 SDK
Docker Desktop or Podman Desktop (required)

Note: Docker Desktop (or Podman) must be running before starting the app. The Intelligence and Dashboard services run in containers.

Setup

# Run the setup script (configures secrets automatically)
./scripts/setup.sh

The script configures LLM provider settings via dotnet user-secrets:

Setting	Description	Default
`llm-provider`	LLM backend: `github`, `azure`, `gemini`, or `openai`	`github`
`github-token`	GitHub PAT for GitHub Models (auto-detected from `gh auth`)	—
`azure-openai-key`	Azure OpenAI API key	—
`azure-openai-endpoint`	Azure OpenAI endpoint URL	—
`google-api-key`	Google Gemini API key	—
`openai-api-key`	OpenAI API key	—
`openai-org-id`	OpenAI organization ID (optional)	—

To use a different provider:

# Azure OpenAI
LLM_PROVIDER=azure AZURE_OPENAI_API_KEY=... AZURE_OPENAI_ENDPOINT=... ./scripts/setup.sh
# Google Gemini
LLM_PROVIDER=gemini GOOGLE_API_KEY=... ./scripts/setup.sh
# OpenAI
LLM_PROVIDER=openai OPENAI_API_KEY=... OPENAI_ORG_ID=... ./scripts/setup.sh

IDE Setup

All IDEs use .NET Aspire to orchestrate services. The AppHost project starts all services (Gateway, Intelligence, Dashboard) with proper dependencies.

VS Code (Recommended)

Install the C# Dev Kit extension
Open the prior-auth folder
Open the Command Palette (Ctrl+Shift+P / Cmd+Shift+P)
Run: ".NET: Open Solution" → select orchestration/AuthScript.sln
In the Solution Explorer, right-click AuthScript.AppHost → "Debug" → "Start New Instance"

Alternatively, from the terminal:

cd orchestration/AuthScript.AppHost
dotnet run

The Aspire Dashboard opens automatically at https://localhost:15888 showing all services.

Visual Studio / Rider

Open orchestration/AuthScript.sln
Set AuthScript.AppHost as the startup project
Press F5 (or Run) to start with debugging
The Aspire Dashboard opens automatically with all services

Schema Synchronization

Regenerate types after modifying API contracts:

npm run sync:schemas

This generates:

TypeScript — React Query hooks, Zod schemas (shared/types/, shared/validation/)
Python — Pydantic models from Gateway spec (apps/intelligence/src/models/generated/)
C# — Records from Intelligence spec (apps/gateway/Gateway.API/Models/Generated/)

CI runs schema sync and fails if generated files drift from committed versions.

Gateway OpenAPI Extraction

Gateway uses runtime OpenAPI generation. To extract locally:

# Start Gateway, then fetch spec
curl http://localhost:5000/openapi/v1.json > apps/gateway/openapi.json
npm run sync:schemas

See shared/schemas/README.md for contract ownership.

Services

Service	Technology	Purpose
Gateway	.NET 10, iText7, Polly	CDS Hooks, FHIR aggregation, PDF generation
Intelligence	Python 3.11, FastAPI	Clinical reasoning (GitHub/Azure/Gemini/OpenAI)
Dashboard	React 19, Vite, TanStack	Shadow dashboard + SMART fallback

Testing

# Run all tests
npm run test# Run specific service tests
npm run test:dashboard
dotnet test apps/gateway/Gateway.API.Tests
# Intelligence tests (run inside container or with local Python)cd apps/intelligence && uv run pytest

Environment Variables

For local development, run ./scripts/setup.sh (see Setup above).

For production or CI/CD, set these environment variables:

Variable	Service	Description
`LLM_PROVIDER`	Intelligence	Provider: `github`, `azure`, `gemini`, `openai`
`GITHUB_TOKEN`	Intelligence	GitHub PAT for GitHub Models
`AZURE_OPENAI_API_KEY`	Intelligence	Azure OpenAI API key
`AZURE_OPENAI_ENDPOINT`	Intelligence	Azure OpenAI endpoint URL
`GOOGLE_API_KEY`	Intelligence	Google Gemini API key
`OPENAI_API_KEY`	Intelligence	OpenAI API key
`OPENAI_ORG_ID`	Intelligence	OpenAI organization ID (optional)
`Epic__ClientId`	Gateway	Epic Launchpad client ID
`Epic__FhirBaseUrl`	Gateway	FHIR R4 endpoint (has default)

License

This is a class project for CSE 589 at the University of Washington.

AuthScript

AI-powered prior authorization automation that integrates directly into Epic's clinical workflow via CDS Hooks. AuthScript automatically aggregates clinical data, reasons against payer policies, and generates completed PA forms.

Documentation

Quick Links	Description
Architecture Overview	System patterns and data flow
Design Document	Detailed technical design

Quick Start

Prerequisites

Node.js 20+ with npm
.NET 10 SDK
Docker Desktop or Podman Desktop (required)

Note: Docker Desktop (or Podman) must be running before starting the app. The Intelligence and Dashboard services run in containers.

Setup

# Run the setup script (configures secrets automatically)
./scripts/setup.sh

The script configures LLM provider settings via dotnet user-secrets:

Setting	Description	Default
`llm-provider`	LLM backend: `github`, `azure`, `gemini`, or `openai`	`github`
`github-token`	GitHub PAT for GitHub Models (auto-detected from `gh auth`)	—
`azure-openai-key`	Azure OpenAI API key	—
`azure-openai-endpoint`	Azure OpenAI endpoint URL	—
`google-api-key`	Google Gemini API key	—
`openai-api-key`	OpenAI API key	—
`openai-org-id`	OpenAI organization ID (optional)	—

To use a different provider:

# Azure OpenAI
LLM_PROVIDER=azure AZURE_OPENAI_API_KEY=... AZURE_OPENAI_ENDPOINT=... ./scripts/setup.sh
# Google Gemini
LLM_PROVIDER=gemini GOOGLE_API_KEY=... ./scripts/setup.sh
# OpenAI
LLM_PROVIDER=openai OPENAI_API_KEY=... OPENAI_ORG_ID=... ./scripts/setup.sh

IDE Setup

All IDEs use .NET Aspire to orchestrate services. The AppHost project starts all services (Gateway, Intelligence, Dashboard) with proper dependencies.

VS Code (Recommended)

Install the C# Dev Kit extension
Open the prior-auth folder
Open the Command Palette (Ctrl+Shift+P / Cmd+Shift+P)
Run: ".NET: Open Solution" → select orchestration/AuthScript.sln
In the Solution Explorer, right-click AuthScript.AppHost → "Debug" → "Start New Instance"

Alternatively, from the terminal:

cd orchestration/AuthScript.AppHost
dotnet run

The Aspire Dashboard opens automatically at https://localhost:15888 showing all services.

Visual Studio / Rider

Open orchestration/AuthScript.sln
Set AuthScript.AppHost as the startup project
Press F5 (or Run) to start with debugging
The Aspire Dashboard opens automatically with all services

Schema Synchronization

Regenerate types after modifying API contracts:

npm run sync:schemas

This generates:

TypeScript — React Query hooks, Zod schemas (shared/types/, shared/validation/)
Python — Pydantic models from Gateway spec (apps/intelligence/src/models/generated/)
C# — Records from Intelligence spec (apps/gateway/Gateway.API/Models/Generated/)

CI runs schema sync and fails if generated files drift from committed versions.

Gateway OpenAPI Extraction

Gateway uses runtime OpenAPI generation. To extract locally:

# Start Gateway, then fetch spec
curl http://localhost:5000/openapi/v1.json > apps/gateway/openapi.json
npm run sync:schemas

See shared/schemas/README.md for contract ownership.

Services

Service	Technology	Purpose
Gateway	.NET 10, iText7, Polly	CDS Hooks, FHIR aggregation, PDF generation
Intelligence	Python 3.11, FastAPI	Clinical reasoning (GitHub/Azure/Gemini/OpenAI)
Dashboard	React 19, Vite, TanStack	Shadow dashboard + SMART fallback

Testing

# Run all tests
npm run test# Run specific service tests
npm run test:dashboard
dotnet test apps/gateway/Gateway.API.Tests
# Intelligence tests (run inside container or with local Python)cd apps/intelligence && uv run pytest

Environment Variables

For local development, run ./scripts/setup.sh (see Setup above).

For production or CI/CD, set these environment variables:

Variable	Service	Description
`LLM_PROVIDER`	Intelligence	Provider: `github`, `azure`, `gemini`, `openai`
`GITHUB_TOKEN`	Intelligence	GitHub PAT for GitHub Models
`AZURE_OPENAI_API_KEY`	Intelligence	Azure OpenAI API key
`AZURE_OPENAI_ENDPOINT`	Intelligence	Azure OpenAI endpoint URL
`GOOGLE_API_KEY`	Intelligence	Google Gemini API key
`OPENAI_API_KEY`	Intelligence	OpenAI API key
`OPENAI_ORG_ID`	Intelligence	OpenAI organization ID (optional)
`Epic__ClientId`	Gateway	Epic Launchpad client ID
`Epic__FhirBaseUrl`	Gateway	FHIR R4 endpoint (has default)

License

This is a class project for CSE 589 at the University of Washington.

AuthScript

AI-powered prior authorization automation that integrates directly into Epic's clinical workflow via CDS Hooks. AuthScript automatically aggregates clinical data, reasons against payer policies, and generates completed PA forms.

Documentation

Quick Links	Description
Architecture Overview	System patterns and data flow
Design Document	Detailed technical design

Quick Start

Prerequisites

Node.js 20+ with npm
.NET 10 SDK
Docker Desktop or Podman Desktop (required)

Note: Docker Desktop (or Podman) must be running before starting the app. The Intelligence and Dashboard services run in containers.

Setup

# Run the setup script (configures secrets automatically)
./scripts/setup.sh

The script configures LLM provider settings via dotnet user-secrets:

Setting	Description	Default
`llm-provider`	LLM backend: `github`, `azure`, `gemini`, or `openai`	`github`
`github-token`	GitHub PAT for GitHub Models (auto-detected from `gh auth`)	—
`azure-openai-key`	Azure OpenAI API key	—
`azure-openai-endpoint`	Azure OpenAI endpoint URL	—
`google-api-key`	Google Gemini API key	—
`openai-api-key`	OpenAI API key	—
`openai-org-id`	OpenAI organization ID (optional)	—

To use a different provider:

# Azure OpenAI
LLM_PROVIDER=azure AZURE_OPENAI_API_KEY=... AZURE_OPENAI_ENDPOINT=... ./scripts/setup.sh
# Google Gemini
LLM_PROVIDER=gemini GOOGLE_API_KEY=... ./scripts/setup.sh
# OpenAI
LLM_PROVIDER=openai OPENAI_API_KEY=... OPENAI_ORG_ID=... ./scripts/setup.sh

IDE Setup

All IDEs use .NET Aspire to orchestrate services. The AppHost project starts all services (Gateway, Intelligence, Dashboard) with proper dependencies.

VS Code (Recommended)

Install the C# Dev Kit extension
Open the prior-auth folder
Open the Command Palette (Ctrl+Shift+P / Cmd+Shift+P)
Run: ".NET: Open Solution" → select orchestration/AuthScript.sln
In the Solution Explorer, right-click AuthScript.AppHost → "Debug" → "Start New Instance"

Alternatively, from the terminal:

cd orchestration/AuthScript.AppHost
dotnet run

The Aspire Dashboard opens automatically at https://localhost:15888 showing all services.

Visual Studio / Rider

Open orchestration/AuthScript.sln
Set AuthScript.AppHost as the startup project
Press F5 (or Run) to start with debugging
The Aspire Dashboard opens automatically with all services

Schema Synchronization

Regenerate types after modifying API contracts:

npm run sync:schemas

This generates:

TypeScript — React Query hooks, Zod schemas (shared/types/, shared/validation/)
Python — Pydantic models from Gateway spec (apps/intelligence/src/models/generated/)
C# — Records from Intelligence spec (apps/gateway/Gateway.API/Models/Generated/)

CI runs schema sync and fails if generated files drift from committed versions.

Gateway OpenAPI Extraction

Gateway uses runtime OpenAPI generation. To extract locally:

# Start Gateway, then fetch spec
curl http://localhost:5000/openapi/v1.json > apps/gateway/openapi.json
npm run sync:schemas

See shared/schemas/README.md for contract ownership.

Services

Service	Technology	Purpose
Gateway	.NET 10, iText7, Polly	CDS Hooks, FHIR aggregation, PDF generation
Intelligence	Python 3.11, FastAPI	Clinical reasoning (GitHub/Azure/Gemini/OpenAI)
Dashboard	React 19, Vite, TanStack	Shadow dashboard + SMART fallback

Testing

# Run all tests
npm run test# Run specific service tests
npm run test:dashboard
dotnet test apps/gateway/Gateway.API.Tests
# Intelligence tests (run inside container or with local Python)cd apps/intelligence && uv run pytest

Environment Variables

For local development, run ./scripts/setup.sh (see Setup above).

For production or CI/CD, set these environment variables:

Variable	Service	Description
`LLM_PROVIDER`	Intelligence	Provider: `github`, `azure`, `gemini`, `openai`
`GITHUB_TOKEN`	Intelligence	GitHub PAT for GitHub Models
`AZURE_OPENAI_API_KEY`	Intelligence	Azure OpenAI API key
`AZURE_OPENAI_ENDPOINT`	Intelligence	Azure OpenAI endpoint URL
`GOOGLE_API_KEY`	Intelligence	Google Gemini API key
`OPENAI_API_KEY`	Intelligence	OpenAI API key
`OPENAI_ORG_ID`	Intelligence	OpenAI organization ID (optional)
`Epic__ClientId`	Gateway	Epic Launchpad client ID
`Epic__FhirBaseUrl`	Gateway	FHIR R4 endpoint (has default)

License

This is a class project for CSE 589 at the University of Washington.

AuthScript

AI-powered prior authorization automation that integrates directly into Epic's clinical workflow via CDS Hooks. AuthScript automatically aggregates clinical data, reasons against payer policies, and generates completed PA forms.

Documentation

Quick Links	Description
Architecture Overview	System patterns and data flow
Design Document	Detailed technical design

Quick Start

Prerequisites

Node.js 20+ with npm
.NET 10 SDK
Docker Desktop or Podman Desktop (required)

Note: Docker Desktop (or Podman) must be running before starting the app. The Intelligence and Dashboard services run in containers.

Setup

# Run the setup script (configures secrets automatically)
./scripts/setup.sh

The script configures LLM provider settings via dotnet user-secrets:

Setting	Description	Default
`llm-provider`	LLM backend: `github`, `azure`, `gemini`, or `openai`	`github`
`github-token`	GitHub PAT for GitHub Models (auto-detected from `gh auth`)	—
`azure-openai-key`	Azure OpenAI API key	—
`azure-openai-endpoint`	Azure OpenAI endpoint URL	—
`google-api-key`	Google Gemini API key	—
`openai-api-key`	OpenAI API key	—
`openai-org-id`	OpenAI organization ID (optional)	—

To use a different provider:

# Azure OpenAI
LLM_PROVIDER=azure AZURE_OPENAI_API_KEY=... AZURE_OPENAI_ENDPOINT=... ./scripts/setup.sh
# Google Gemini
LLM_PROVIDER=gemini GOOGLE_API_KEY=... ./scripts/setup.sh
# OpenAI
LLM_PROVIDER=openai OPENAI_API_KEY=... OPENAI_ORG_ID=... ./scripts/setup.sh

IDE Setup

All IDEs use .NET Aspire to orchestrate services. The AppHost project starts all services (Gateway, Intelligence, Dashboard) with proper dependencies.

VS Code (Recommended)

Install the C# Dev Kit extension
Open the prior-auth folder
Open the Command Palette (Ctrl+Shift+P / Cmd+Shift+P)
Run: ".NET: Open Solution" → select orchestration/AuthScript.sln
In the Solution Explorer, right-click AuthScript.AppHost → "Debug" → "Start New Instance"

Alternatively, from the terminal:

cd orchestration/AuthScript.AppHost
dotnet run

The Aspire Dashboard opens automatically at https://localhost:15888 showing all services.

Visual Studio / Rider

Open orchestration/AuthScript.sln
Set AuthScript.AppHost as the startup project
Press F5 (or Run) to start with debugging
The Aspire Dashboard opens automatically with all services

Schema Synchronization

Regenerate types after modifying API contracts:

npm run sync:schemas

This generates:

TypeScript — React Query hooks, Zod schemas (shared/types/, shared/validation/)
Python — Pydantic models from Gateway spec (apps/intelligence/src/models/generated/)
C# — Records from Intelligence spec (apps/gateway/Gateway.API/Models/Generated/)

CI runs schema sync and fails if generated files drift from committed versions.

Gateway OpenAPI Extraction

Gateway uses runtime OpenAPI generation. To extract locally:

# Start Gateway, then fetch spec
curl http://localhost:5000/openapi/v1.json > apps/gateway/openapi.json
npm run sync:schemas

See shared/schemas/README.md for contract ownership.

Services

Service	Technology	Purpose
Gateway	.NET 10, iText7, Polly	CDS Hooks, FHIR aggregation, PDF generation
Intelligence	Python 3.11, FastAPI	Clinical reasoning (GitHub/Azure/Gemini/OpenAI)
Dashboard	React 19, Vite, TanStack	Shadow dashboard + SMART fallback

Testing

# Run all tests
npm run test# Run specific service tests
npm run test:dashboard
dotnet test apps/gateway/Gateway.API.Tests
# Intelligence tests (run inside container or with local Python)cd apps/intelligence && uv run pytest

Environment Variables

For local development, run ./scripts/setup.sh (see Setup above).

For production or CI/CD, set these environment variables:

Variable	Service	Description
`LLM_PROVIDER`	Intelligence	Provider: `github`, `azure`, `gemini`, `openai`
`GITHUB_TOKEN`	Intelligence	GitHub PAT for GitHub Models
`AZURE_OPENAI_API_KEY`	Intelligence	Azure OpenAI API key
`AZURE_OPENAI_ENDPOINT`	Intelligence	Azure OpenAI endpoint URL
`GOOGLE_API_KEY`	Intelligence	Google Gemini API key
`OPENAI_API_KEY`	Intelligence	OpenAI API key
`OPENAI_ORG_ID`	Intelligence	OpenAI organization ID (optional)
`Epic__ClientId`	Gateway	Epic Launchpad client ID
`Epic__FhirBaseUrl`	Gateway	FHIR R4 endpoint (has default)

License

This is a class project for CSE 589 at the University of Washington.

AuthScript

AI-powered prior authorization automation that integrates directly into Epic's clinical workflow via CDS Hooks. AuthScript automatically aggregates clinical data, reasons against payer policies, and generates completed PA forms.

Documentation

Quick Links	Description
Architecture Overview	System patterns and data flow
Design Document	Detailed technical design

Quick Start

Prerequisites

Node.js 20+ with npm
.NET 10 SDK
Docker Desktop or Podman Desktop (required)

Note: Docker Desktop (or Podman) must be running before starting the app. The Intelligence and Dashboard services run in containers.

Setup

# Run the setup script (configures secrets automatically)
./scripts/setup.sh

The script configures LLM provider settings via dotnet user-secrets:

Setting	Description	Default
`llm-provider`	LLM backend: `github`, `azure`, `gemini`, or `openai`	`github`
`github-token`	GitHub PAT for GitHub Models (auto-detected from `gh auth`)	—
`azure-openai-key`	Azure OpenAI API key	—
`azure-openai-endpoint`	Azure OpenAI endpoint URL	—
`google-api-key`	Google Gemini API key	—
`openai-api-key`	OpenAI API key	—
`openai-org-id`	OpenAI organization ID (optional)	—

To use a different provider:

# Azure OpenAI
LLM_PROVIDER=azure AZURE_OPENAI_API_KEY=... AZURE_OPENAI_ENDPOINT=... ./scripts/setup.sh
# Google Gemini
LLM_PROVIDER=gemini GOOGLE_API_KEY=... ./scripts/setup.sh
# OpenAI
LLM_PROVIDER=openai OPENAI_API_KEY=... OPENAI_ORG_ID=... ./scripts/setup.sh

IDE Setup

All IDEs use .NET Aspire to orchestrate services. The AppHost project starts all services (Gateway, Intelligence, Dashboard) with proper dependencies.

VS Code (Recommended)

Install the C# Dev Kit extension
Open the prior-auth folder
Open the Command Palette (Ctrl+Shift+P / Cmd+Shift+P)
Run: ".NET: Open Solution" → select orchestration/AuthScript.sln
In the Solution Explorer, right-click AuthScript.AppHost → "Debug" → "Start New Instance"

Alternatively, from the terminal:

cd orchestration/AuthScript.AppHost
dotnet run

The Aspire Dashboard opens automatically at https://localhost:15888 showing all services.

Visual Studio / Rider

Open orchestration/AuthScript.sln
Set AuthScript.AppHost as the startup project
Press F5 (or Run) to start with debugging
The Aspire Dashboard opens automatically with all services

Schema Synchronization

Regenerate types after modifying API contracts:

npm run sync:schemas

This generates:

TypeScript — React Query hooks, Zod schemas (shared/types/, shared/validation/)
Python — Pydantic models from Gateway spec (apps/intelligence/src/models/generated/)
C# — Records from Intelligence spec (apps/gateway/Gateway.API/Models/Generated/)

CI runs schema sync and fails if generated files drift from committed versions.

Gateway OpenAPI Extraction

Gateway uses runtime OpenAPI generation. To extract locally:

# Start Gateway, then fetch spec
curl http://localhost:5000/openapi/v1.json > apps/gateway/openapi.json
npm run sync:schemas

See shared/schemas/README.md for contract ownership.

Services

Service	Technology	Purpose
Gateway	.NET 10, iText7, Polly	CDS Hooks, FHIR aggregation, PDF generation
Intelligence	Python 3.11, FastAPI	Clinical reasoning (GitHub/Azure/Gemini/OpenAI)
Dashboard	React 19, Vite, TanStack	Shadow dashboard + SMART fallback

Testing

# Run all tests
npm run test# Run specific service tests
npm run test:dashboard
dotnet test apps/gateway/Gateway.API.Tests
# Intelligence tests (run inside container or with local Python)cd apps/intelligence && uv run pytest

Environment Variables

For local development, run ./scripts/setup.sh (see Setup above).

For production or CI/CD, set these environment variables:

Variable	Service	Description
`LLM_PROVIDER`	Intelligence	Provider: `github`, `azure`, `gemini`, `openai`
`GITHUB_TOKEN`	Intelligence	GitHub PAT for GitHub Models
`AZURE_OPENAI_API_KEY`	Intelligence	Azure OpenAI API key
`AZURE_OPENAI_ENDPOINT`	Intelligence	Azure OpenAI endpoint URL
`GOOGLE_API_KEY`	Intelligence	Google Gemini API key
`OPENAI_API_KEY`	Intelligence	OpenAI API key
`OPENAI_ORG_ID`	Intelligence	OpenAI organization ID (optional)
`Epic__ClientId`	Gateway	Epic Launchpad client ID
`Epic__FhirBaseUrl`	Gateway	FHIR R4 endpoint (has default)

License

This is a class project for CSE 589 at the University of Washington.

Name		Name	Last commit message	Last commit date
.config		.config
.github		.github
apps		apps
assets/pdf-templates		assets/pdf-templates
docs		docs
orchestration		orchestration
scripts		scripts
shared		shared
.coderabbit.yaml		.coderabbit.yaml
.dockerignore		.dockerignore
.gitignore		.gitignore
.node-version		.node-version
README.md		README.md
orval.config.ts		orval.config.ts
package-lock.json		package-lock.json
package.json		package.json

Name		Name	Last commit message	Last commit date
.config		.config
.github		.github
apps		apps
assets/pdf-templates		assets/pdf-templates
docs		docs
orchestration		orchestration
scripts		scripts
shared		shared
.coderabbit.yaml		.coderabbit.yaml
.dockerignore		.dockerignore
.gitignore		.gitignore
.node-version		.node-version
README.md		README.md
orval.config.ts		orval.config.ts
package-lock.json		package-lock.json
package.json		package.json

Name		Name	Last commit message	Last commit date
.config		.config
.github		.github
apps		apps
assets/pdf-templates		assets/pdf-templates
docs		docs
orchestration		orchestration
scripts		scripts
shared		shared
.coderabbit.yaml		.coderabbit.yaml
.dockerignore		.dockerignore
.gitignore		.gitignore
.node-version		.node-version
README.md		README.md
orval.config.ts		orval.config.ts
package-lock.json		package-lock.json
package.json		package.json

Name		Name	Last commit message	Last commit date
.config		.config
.github		.github
apps		apps
assets/pdf-templates		assets/pdf-templates
docs		docs
orchestration		orchestration
scripts		scripts
shared		shared
.coderabbit.yaml		.coderabbit.yaml
.dockerignore		.dockerignore
.gitignore		.gitignore
.node-version		.node-version
README.md		README.md
orval.config.ts		orval.config.ts
package-lock.json		package-lock.json
package.json		package.json

Name		Name	Last commit message	Last commit date
.config		.config
.github		.github
apps		apps
assets/pdf-templates		assets/pdf-templates
docs		docs
orchestration		orchestration
scripts		scripts
shared		shared
.coderabbit.yaml		.coderabbit.yaml
.dockerignore		.dockerignore
.gitignore		.gitignore
.node-version		.node-version
README.md		README.md
orval.config.ts		orval.config.ts
package-lock.json		package-lock.json
package.json		package.json

Uh oh!

Latest commit

History

Folders and files

Repository files navigation

AuthScript

Documentation

Quick Start

Prerequisites

Setup

IDE Setup

VS Code (Recommended)

Visual Studio / Rider

Schema Synchronization

Gateway OpenAPI Extraction

Services

Testing

Environment Variables

License

About

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages

Uh oh!

Latest commit

History

Folders and files

Repository files navigation

AuthScript

Documentation

Quick Start

Prerequisites

Setup

IDE Setup

VS Code (Recommended)

Visual Studio / Rider

Schema Synchronization

Gateway OpenAPI Extraction

Services

Testing

Environment Variables

License

About

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages

Uh oh!

Latest commit

History

Folders and files

Repository files navigation

AuthScript

Documentation

Quick Start

Prerequisites

Setup

IDE Setup

VS Code (Recommended)

Visual Studio / Rider

Schema Synchronization

Gateway OpenAPI Extraction

Services

Testing

Environment Variables

License

About

Resources

Stars

Watchers

Forks