AI Pipelines with Telegram Bot

This project provides a framework for creating and executing AI-powered pipelines, with a Telegram bot interface for easy interaction.

Features

• Execute AI-powered pipelines using OpenAI's models
• Integration with Model Context Protocol (MCP) tools
• Telegram bot interface for easy interaction
• Support for dynamic prompts and variables

Prerequisites

• Node.js and Bun installed
• OpenAI API key
• Telegram Bot Token (get it from @BotFather)

Setup

1. Clone the repository
2. Install dependencies:

bun install

3. Copy .env.example to .env and fill in your credentials:

cp .env.example .env

4. Edit the .env file and add your:
   • OPENAI_API_KEY - Your OpenAI API key
   • TELEGRAM_BOT_TOKEN - Your Telegram bot token from @BotFather
   • MCP_SERVER_URL - (Optional) MCP server URL (defaults to Santiment's MCP server)

Running the Bot

Start the bot with:

bun start

Or in development mode with auto-reload:

bun --watch src/main.ts

Using the Bot

1. Find your bot on Telegram using the username you set with @BotFather
2. Start a chat with the bot
3. Send any message to the bot, and it will process it through the AI pipeline
4. The bot will respond with the AI's analysis or action

Project Structure

• src/bot/ - Telegram bot implementation
• src/models/ - Data models and interfaces
• src/pipelines/ - Pipeline execution logic
• src/services/ - External service integrations
• src/main.ts - Application entry point

Development

• Run linter: bun run lint
• Run type checking: bun run ts:check
• Run both: bun run check

License

MIT