Veris Claims Eligibility

OpenAI-backed implementation of the medical bill eligibility analyzer handoff. The app lets a reviewer load one or more bill documents, extract claim fields with an OpenAI model, edit extracted values, tune plain-English eligibility rules, and export a CSV that opens cleanly in Excel.

Screenshots

Run locally

1. Open PowerShell.

2. Go to the project directory:

   cd "C:\Users\nafer\github repo\Veris Claims Eligibility"

3. Install dependencies:

   npm install

4. Start the backend:

   npm start

5. Open this URL in your browser:

   http://localhost:4173
   

The backend reads OpenAI settings from .env. Keep .env local; it is ignored by git.

Use this shape for the model menu:

OPENAI_API_KEY=your-secret-key
OPENAI_MODEL=gpt-5.6-luna
OPENAI_MODEL_OPTIONS=gpt-5.6-luna|Efficient - high volume;gpt-5.5|Highest accuracy - slower;gpt-5.4-mini|Fast - high accuracy;gpt-5.4-nano|Lowest cost - fastest

Each model option is model-id|Description, and options are separated with semicolons. The UI displays the model ID as the label. OPENAI_MODEL must match one of the option IDs to choose the default selected model.

Included behavior

• Server-side OpenAI extraction using the configured .env model.
• Drag/drop and file picker support for PDF, JPG, and PNG bill documents.
• Each uploaded file is analyzed as a separate bill document.
• Multi-page PDFs can be reviewed in the source document panel.
• Editable extracted fields with missing-required-field highlighting.
• Plain-English rule parser for required fields and excluded service categories.
• Flagged/eligible filters.
• CSV export named eligibility-results.csv.

License

MIT. See LICENSE.