Project README

Overview

This is a personal repo for my work at a startup. The architecture leverages AWS services, Deepgram, OpenAI, Twilio, and Plivo to handle outbound calls, speech-to-text, text-to-speech, and real-time interactions.

Architecture

Below is the architecture diagram for the project:

![Architecture Diagram]

Environment Variables

Create a .env file based on the following template:

# Twilio Main Account
TWILIO_ACCOUNT_SID=
TWILIO_AUTH_TOKEN=
TWILIO_PHONE_NUMBER=

# Deepgram Keys
DEEPGRAM_API_KEY1=
DEEPGRAM_API_KEY2=
DEEPGRAM_API_KEY3=
DEEPGRAM_API_KEY4=
DEEPGRAM_API_KEY5=
DEEPGRAM_API_KEY6=

# Open AI
OPENAI_API_KEY=

# AWS Credentials
AWS_ACCESS_KEY_ID=
AWS_SECRET_ACCESS_KEY=
AWS_REGION=

PORT=8080

Notes

• For Deepgram API keys, you can use only one. Multiple keys are provided here to handle more calls since Deepgram has a 100 requests limit for the free version.

Installation & Run

1. Install dependencies:

   npm i

2. Build the project:

   npm run build

3. Start the project:

   npm run start