🤖 Gemini AI Telegram Bot

A powerful Telegram bot powered by Google's Gemini 2.0 Flash for natural conversations and stunning AI image generation with multimodal capabilities.

• Advanced Image Generation: Create stunning images using Gemini 2.0 Flash's native image generation capabilities
  • Consistent character and setting representation across multiple generated images
  • Superior text rendering in generated images
  • Conversational image editing through natural dialogue
  • Accurate representation of complex concepts using world knowledge
• Natural Conversation: Engage in fluid, context-aware conversations with Gemini AI
• Image Analysis: Send any image for AI-powered analysis and insights
• Smart History: Maintains chat history for more contextual responses
• True Multi-Modal Interaction: Seamlessly process text and images in the same conversation

Generate consistent visual narratives where characters and settings remain cohesive throughout multiple images.

Fine-tune generated images through natural conversation, allowing for precise adjustments and creative exploration.

Create detailed, accurate images backed by Gemini's understanding of real-world concepts - perfect for technical illustrations or educational content.

Generate images with crisp, accurate text rendering - ideal for creating:

• Social media posts
• Advertisements
• Event invitations
• Infographics

• Node.js (v18 or newer)
• A Telegram Bot Token from @BotFather
• A Google Gemini API Key from Google AI Studio

1. Clone & Install

   git clone https://github.com/jiaweing/GeminiTelegramBot
   cd GeminiTelegramBot
   pnpm install

2. Configure Environment

   cp .env.example .env

   Edit .env with your tokens:

   BOT_TOKEN=your_telegram_bot_token
   GEMINI_API_KEY=your_gemini_api_key

3. Start the Bot

   pnpm dev

• /start - Initialize or reset the bot
• /help - View usage instructions
• /clear - Clear conversation history
• Type clear - Alternative way to reset chat

• Storytelling

  generate: Tell a story about a cute baby turtle's adventure with illustrations in 3D digital art style
  

• Iterative Design

  generate: A cozy coffee shop interior
  [Then refine with]: Make it more modern with large windows
  [And further]: Add some plants and wooden elements
  

• Technical Visualization

  generate: Step-by-step illustration of how a solar panel works
  

• Text-Rich Images

  generate: Create a modern event poster for a tech conference with the title "Future of AI 2025" and all the important details
  

• Send any image with prompts like:
  • "What's unique about this architectural style?"
  • "Can you analyze the composition of this photograph?"
  • "What emotions does this artwork convey?"

• "Explain quantum computing with visual examples"
• "Design a logo for an eco-friendly startup"
• "Create a visual story about space exploration"

graph TD
    A[User Input] --> B[Telegram Bot]
    B --> C{Input Type}
    C -->|Text| D[Text Processing]
    C -->|Image| E[Image Analysis]
    C -->|Generate| F[Gemini 2.0 Flash]
    D --> G[Gemini AI]
    E --> G
    F --> G
    G --> H[Response Handler]
    H --> I[Chat History]
    H --> J[File System]
    H --> K[User Response]

• Framework: Telegraf for Telegram Bot API
• AI Engine: @google/genai for Gemini integration
  • Uses Gemini 2.0 Flash experimental model for enhanced image generation
• Language: TypeScript for type safety and better development experience
• Runtime: Node.js with support for modern ES modules

The bot includes comprehensive error handling for:

• Network connectivity issues
• API rate limits and timeouts
• Invalid input formats
• File system operations
• AI processing errors

Contributions are welcome! Please feel free to submit a Pull Request.

MIT