CharlesBot

CharlesBot is a Telegram bot that leverages the OpenAI API to chat with users and generate images using the DALL-E model. You can start a conversation by sending a text message or requesting an image by sending a message starting with "Generate image:" followed by the image description.

Requirements

• Python 3.6+
• aiogram
• openai

Installation

1. Clone the repository:

git clone https://github.com/Kryohi/GPT_TG-Bot.git
cd GPT_TG-Bot

2. Create a virtual environment and activate it:

python -m venv venv
source venv/bin/activate    # On Windows use 'venv\Scripts\activate'

3. Install the required packages:

pip install --user aiogram openai pydub

4. Set up the environment variables with your OpenAI API key, Telegram Bot API key, and your Telegram user ID:

export OPENAI_API_KEY="your_openai_api_key"
export TELEGRAM_API_KEY_CHATBOT="your_telegram_bot_api_key"
export TG_USERID="your_telegram_user_id"

# On Windows use 'set' instead of 'export':
# set OPENAI_API_KEY="your_openai_api_key"
# set TELEGRAM_API_KEY_CHATBOT="your_telegram_bot_api_key"
# set TG_USERID="your_telegram_user_id"

If you don't know how to create a Telegram bot and obtain its API key, follow the instructions in the Telegram BotFather documentation. For the OpenAI API keys: https://platform.openai.com/account/api-keys.

Usage

To start the bot, run:

python main.py

To interact with the bot, send a message to your bot on Telegram or add the bot to a group. The bot will respond to any message that mentions it or is a direct reply to one of its messages.

To request an image, send a message starting with "Generate image:" followed by the description. For instance:

Generate image: a beautiful mountain landscape

The bot will generate and send back an image based on the description.

If you answer to one of its messages with a voice message, it will answer as if you had written it.

Future Development

• Use a different model to generate images (Stable Diffusion?)
• Fix audio transcription.
• Access to the web.
• PineCone integration.

License

This project is licensed under the terms of the MIT License.