InterviewGPT

WIP

Your personal interview assistant companion

Setup

Setting up Development:

Assuming you have Python 3.9+, and have installed poetry

• git clone https://github.com/ericz99/interview-gpt.git <desired dest>

• Run poetry install to install all dependency

• Initialize environment poetry shell

• How to run?

  • You can run with web / server

    • Navigate both /backend/interview_gpt/server and /web
      • On backend run uvicorn main:app --reload
      • On web run pnpm run dev

  • Or just run the core feature

    • Navigate both /backend/interview_gpt/core
      • Run poetry run python main.py

API Key

• .env file:
  • Create a copy of .env.template named .env
  • Add your OPENAI_API_KEY in .env

Development

• This is currently project not optimized for performance as of now.
• Working on writing a simple api to create new instance of recorders, and new session
• trying to add VAD (Voice activity detection) instead of relying on RMS
• currently this can only spin up one instance of recorder, and it uses socketio to send data from backend to client
• sometime theres like a delay when speaking to the mic, then couple second later it process the audio data

tldr: not slow nor fast, but works good and transcription is on point. Needs optimized using VAD (Voice activity detection)

Usage

After running the script, the program will start listening to system audio (eg: Speaker), and will record + create .wav file + transcript for you. After, the AI will read your .wav to create a transcript then will help you answer question or response back about the transcript.