Insurance LiveKit Agent

A conversational voice agent that collects insurance information from patients, performs STEDI eligibility checks, and communicates patient copay information for office visits.

Project Overview

This application uses LiveKit Workflows to create a voice agent that:

1. Collects patient information (first name, last name, insurance member ID)
2. Sends this information to the STEDI eligibility API
3. Analyzes the insurance response to determine:
   • If the insurance is active
   • If STC 98 (office visits) is covered
   • If the provider is in-network
   • What the copay amount is
4. Communicates the results to the patient in a natural, conversational manner

Prerequisites

• Python 3.10+
• LiveKit account
• STEDI API key
• ElevenLabs API key
• Deepgram API key
• OpenAI API key

Installation

1. Clone this repository
2. Install the required dependencies:

pip install -r requirements.txt

3. Create a .env file in the project root with the following variables:

LIVEKIT_API_KEY=your_livekit_api_key
LIVEKIT_API_SECRET=your_livekit_api_secret
OPENAI_API_KEY=your_openai_api_key
ELEVENLABS_API_KEY=your_elevenlabs_api_key
DEEPGRAM_API_KEY=your_deepgram_api_key
STEDI_API_KEY=your_stedi_api_key

Running the Application

To run the Insurance LiveKit Agent:

python main.py

This will start the LiveKit agent which will:

• Initialize the necessary plugins (OpenAI, ElevenLabs, Deepgram)
• Set up the conversation workflow
• Begin the patient interaction process

Project Structure

• main.py: Entry point for the application
• models.py: Data models for insurance session information
• stedi_api.py: Integration with STEDI eligibility API
• utils.py: Utility functions
• agents/: Directory containing agent implementations
  • collect_info_agent.py: Agent that collects patient information
  • eligibility_agent.py: Agent that processes eligibility results