This Telegram bot uses OpenAI's GPT-4o-mini model via the AI SDK to provide intelligent responses to user messages.

• Responds to text messages using OpenAI's GPT-4o-mini
• Shows typing indicator while generating responses
• Handles errors gracefully
• Concise and friendly AI responses optimized for chat

1. TELEGRAM_TOKEN: Your Telegram bot token

   • Get this from @BotFather on Telegram
   • Create a new bot with /newbot command
   • Copy the token provided

2. OPENAI_API_KEY: Your OpenAI API key

   • Get this from OpenAI Platform
   • Create a new API key
   • Make sure you have credits in your OpenAI account

• /start - Welcome message and bot introduction
• Any text message - Gets an AI-generated response

The bot's personality and behavior can be customized by editing the system-prompt.md file. This markdown file contains:

• The main system prompt that defines the AI's role and behavior
• Guidelines for response style and tone
• Instructions for handling specific scenarios

Simply edit the markdown file and redeploy to update the bot's behavior. The system includes a fallback prompt in case the file cannot be loaded.

• Uses Grammy framework for Telegram bot functionality
• Integrates AI SDK with OpenAI provider
• Configured to use GPT-4o-mini model (cost-effective and fast)
• Maximum response length: 500 tokens
• System prompt loaded from system-prompt.md file for easy customization

• Validates required environment variables on startup
• Handles API errors gracefully with user-friendly messages
• Logs errors for debugging
• Only responds to text messages (ignores media, stickers, etc.)

1. Set up the required environment variables in Val Town
2. Customize the system prompt by editing system-prompt.md
3. Deploy the val as an HTTP endpoint
4. Start chatting with your bot on Telegram!

The bot will automatically set up the webhook when it receives its first request.