Gemini API Command Line Chat Bot

chat.py

import os
import json
from datetime import datetime
from dotenv import load_dotenv
from colorama import Fore, Back, Style

load_dotenv()

import google.generativeai as genai

genai.configure(api_key=os.environ["GOOGLE_API_KEY"])
model = genai.GenerativeModel("gemini-1.5-pro-latest")
message_history = []

def get_response(chat, new_question):

    response = chat.send_message(new_question)
    for chunk in response:
      print(chunk.text)
    
    return response

def main():
    os.system("cls" if os.name == "nt" else "clear")
    previous_questions_and_answers = []
    while True:

        try:
          chat = model.start_chat(history=[])
          new_question = input(
              Fore.BLUE + Style.BRIGHT + "Chat: " + Style.RESET_ALL
          )

          if new_question:
            response = get_response(chat, new_question)          
          else:
            continue
          message_history.append(response.text)
        
        except KeyboardInterrupt:
          
          print("\nGoodbye!")

          now = datetime.now().strftime("%Y-%m-%d_%H-%M-%S")
          filename = f"chat-{now}.json"
          with open(filename, 'w') as f:
            json.dump(message_history, f, indent=6)
            
          message_history.clear()
          break

if __name__ == "__main__":
    main()

Important

You need an API key from Google to operate the above code, you can find one here.

Once you have an API key, setup a .env file with the following text in the same folder as your code.

GOOGLE_API_KEY = "<YOUR API KEY>"

Text is styled, but this is optional; you can disable styling by replacing line 31 with "Chat: ". Additionally, all chat history is automatically stored as local JSON; you can disable this by commenting out lines 44-47. For both disables, make sure to remove unused all imports.

---

A simple command line chat bot that uses Google's Gemini API.

This works as a free alternative to any CLI chat bots using OpenAI, Anthropic, etc.

You can learn more about the Gemini API at https://ai.google.dev/