This tutorial explains step-by-step how to build a custom Telegram chatbot that can interact with OpenAI ChatGPT. The tutorial is written in Python and uses the python-telegram-bot and openai packages.
1) Create Telegram Chatbot
Open your IDE and create a file named telegram-bot.py
We are going to use this https://github.com/python-telegram-bot/python-telegram-bot package that will help us to create the telegram bot. Make sure to install it with
pip3 install python-telegram-bot
After installing, paste this code to your telegram-bot.py file:
import logging
import os
from telegram import Update
from telegram.ext import (ApplicationBuilder, CommandHandler, ContextTypes,
MessageHandler, filters)
logging.basicConfig(
format='%(asctime)s - %(name)s - %(levelname)s - %(message)s',
level=logging.INFO
)
TELEGRAM_API_TOKEN = os.getenv("TELEGRAM_API_TOKEN")
async def start(update: Update, context: ContextTypes.DEFAULT_TYPE):
await context.bot.send_message(chat_id=update.effective_chat.id, text="I'm a bot, please talk to me!")
async def echo(update: Update, context: ContextTypes.DEFAULT_TYPE):
await context.bot.send_message(chat_id=update.effective_chat.id, text=update.message.text)
if __name__ == '__main__':
application = ApplicationBuilder().token(TELEGRAM_API_TOKEN).build()
start_handler = CommandHandler('start', start)
echo_handler = MessageHandler(filters.TEXT & (~filters.COMMAND), echo)
application.add_handler(start_handler)
application.add_handler(echo_handler)
application.run_polling()
We are defining two handlers, one command helper (so if you type /start in telegram), and one message handler.
The message handler (function echo) for testing purposes only gives back what the user typed. So let's try out our bot. But before we can do this we need to register our bot at telegram in order to get the API_KEY.
1.1) Register our bot at telegram
Head over to https://web.telegram.org/k/ and log in with your smartphone. Use the search bar to search for “BotFather”.
Start the registration with the /start command.
It lists all possible commands, we go further with /newbot to register a new bot.
Now try to find a good name for your bot, I had some struggles:
After successful registration, you will get a message from BotFather that your bot is registered, where you can find it, and the token to access it. Copy this token
Create a .env file in the same directory as the telegram-bot.py file, and add the following line to it:
TELEGRAM_API_TOKEN=<your_telegram_api_token>
Replace <your_telegram_api_token> with your actual Telegram API token provided by BotFather.
We can now test your telegram bot! In your IDE open a terminal and in the path where your script is start the bot with
python3 telegram-bot.py
Open the link where you can find your bot or use the search bar on Telegram to find it. Click the start button and let's write something.
Sweet! We have successfully achieved part 1. What we now need to do is to connect our bot with ChatGPT. The idea is simple, the bot is taking the user's input. We will query ChatGPT with this input and redirect the answer from ChatGPT back to our user. Let's go!
2) Create the CHAT-GPT Client
Let’s create a new Python file that will handle all the ChatGPT Stuff, I will call it
“chatgpt_client.py”.
We are going to use the official openai python package to query chatgpt api. lets install the openai python package with
pip3 install openai
Copy this code to the new created file:
import os
import openai
openai.api_key = os.getenv("OPENAI_API_KEY")
def request_chat_gpt(user_message):
try:
completion = openai.ChatCompletion.create(
model="gpt-3.5-turbo",
messages=[
{"role": "user", "content": user_message}
]
)
return completion.choices[0].message['content']
except Exception as e:
print(f"An error occurred: {e}")
return "" # Return an empty string or handle the error appropriately
Now move to the openAI website and get your token.
Place the API key in the .env file, add the following line:
OPENAI_API_KEY="<your_openai_api_key>"
Replace <your_openai_api_key> with your actual OpenAI API key.
To use the openai integration in the echo handler, we can replace the line await context.bot.send_message(chat_id=update.effective_chat.id, text=update.message.text) with the following:
response = request_chat_gpt(update.message.text)
await context.bot.send_message(chat_id=update.effective_chat.id, text=response)
Of course, don’t forget to import the function
from chatgpt_client import request_chat_gpt
This will send the user's input to the request_chat_gpt function, which will use the OpenAI API to generate a response. The response will then be sent back to the user via the echo handler.
Thats all! I hope you enjoyed this one. Next time we are going to build an AirBnB Chatbot that can respond to guests with custom knowledge like Wifi Passwort or Checkout Times, sounds cool right?
If you don't want to miss this article go to my site and subscribe to the Newsletter, I would appreciate this ❤️!