When ChatGPT was first released, people joked that maybe ChatGPT was just some guy on the other end of the computer furiously typing away to answer questions. We thought it would be funny to make this a reality!

KindaGPT is a web application where users can either:

1. Act as a user, interacting with the app just like they would with ChatGPT—except the responses are human-generated by another person using the application.
2. Act as the LLM, where they attempt to generate responses like an AI.

On the LLM side, users receive actual AI-generated feedback on their responses to help them improve as an LLM imposter! Think reinforcement learning with human feedback (RLHF)—except in reverse!

• Live Messaging: WebSockets
• Frontend: React + Next.js
• AI Feedback: OpenAI API

1. Set Up Your API Key

   • You need an OpenAI API key with a funded account.
   • Create a .env.local file in /chatgpt-clone and add your key under the variable name:

     OPENAI_API_KEY=your_api_key_here
     

2. Start the Backend

   • Navigate to the backend directory:

     cd backend

   • Install dependencies:

     pip install -r requirements.txt

   • Run the server:

     python app.py

3. Start the Frontend

   • Open another terminal and navigate to the frontend directory:

     cd chatgpt-clone

   • Install dependencies:

     npm install --force

   • Run the client:

     npm run dev

4. Test It Out!

   • Run one instance of the client as the user and another as the LLM to experience the interaction.