Telegram Bot with AI Memory & Persona Management

A sophisticated Telegram bot that uses AI personas, memory management, and multiple AI services to create engaging, context-aware conversations in group chats.

🚀 Features

• AI-Powered Personas: Multiple distinct AI personalities with unique voices and expertise
• Memory Integration: Remembers past conversations using Mem0 for contextual responses
• Multi-Service AI: Combines OpenAI GPT, Grok, and other AI services
• Real-time Query Handling: Processes fact-based queries with up-to-date information
• Intelligent Reactions: Context-aware responses to group messages
• Topic Initiation: Proactively starts engaging conversations
• Persona Stickiness: Maintains character consistency across conversations
• Firestore Integration: Persistent message storage and retrieval

🏗️ Architecture

src/
├── core_logic/
│   ├── response_logic.py    # Main response handling logic
│   ├── memory.py           # Memory management with Mem0
│   └── llm_personas.py     # Persona management system
├── services/
│   ├── openai_chat.py      # OpenAI API integration
│   ├── grok_chat.py        # Grok API integration
│   ├── fetch_db.py         # Database operations
│   └── state_manager.py    # State management
├── workers/
│   ├── brain.py            # Core processing worker
│   └── sender.py           # Message sending worker
└── config/
    └── settings.py         # Configuration management

🛠️ Setup

Prerequisites

• Python 3.8+
• Telegram Bot Token
• OpenAI API Key
• Grok API Key
• Mem0 API Key
• Google Cloud Firestore credentials

Installation

1. Clone the repository

git clone <repository-url>
cd telegram_bot_tg

2. Install dependencies

pip install -r requirements.txt

3. Environment Setup Create a .env file in the root directory:

# Telegram
TELEGRAM_BOT_TOKEN=your_telegram_bot_token
TELEGRAM_GROUP_ID=your_group_id

# AI Services
OPENAI_API_KEY=your_openai_api_key
GROK_API_KEY=your_grok_api_key
MEM0_API_KEY=your_mem0_api_key

# Google Cloud (for Firestore)
GOOGLE_APPLICATION_CREDENTIALS=path/to/your/service-account.json

4. Configure Settings Update config/settings.py with your specific configuration:

APP_CONFIG = {
    'telegram_group_id': 'your_group_id',
    'sender_bot_users': ['bot_username'],
    'response_context_messages': 10,
    # ... other settings
}

5. Setup Personas Create persona embeddings for better matching:

python -c "from src.core_logic.llm_personas import PersonaManager; pm = PersonaManager(); pm.generate_embeddings()"

🎭 Persona System

The bot includes multiple AI personas, each with unique characteristics:

• Crypto OG: Experienced blockchain enthusiast
• DeFi Degen: High-risk DeFi trader
• NFT Collector: Digital art and collectibles expert
• Tech Minimalist: Privacy and security focused
• Community Builder: Engagement and networking specialist
• Meme Lord: Humor and viral content creator
• Data Analyst: Numbers and analytics focused
• Venture Capitalist: Investment and funding expert
• Developer: Technical implementation specialist
• Enthusiast: General crypto enthusiast

Each persona has:

• Unique voice and tone
• Specialized expertise areas
• Distinct communication patterns
• Emoji usage preferences

🧠 Memory System

The bot uses Mem0 for advanced memory management:

Memory Functions

• get_memory_context(query): Retrieves relevant past interactions
• add_to_memory(content, role): Stores new interactions
• handle_memory(query, type): Handles both query and response storage

Memory Types

• Query Memory: Searches for relevant context before responding
• Response Memory: Stores bot responses for future reference
• Context Integration: Automatically adds memory context to all AI prompts

🔄 Core Logic Flow

1. Message Triage

# Messages are categorized into:
- PERSONA_OPINION: Regular group interactions
- REALTIME_QUERY: Fact-based queries
- TOPIC_INITIATION: Conversation starters

2. Response Generation

# Three main handlers:
- handle_reaction(): Persona-based responses
- handle_realtime_query(): Fact gathering + humanization
- handle_initiation(): Topic generation

3. Memory Integration

# Before each AI call:
memory_context = get_memory_context(message)
prompt = f"##0. Previous chat Context: {memory_context}\n{main_prompt}"

# After response generation:
add_to_memory(user_message, "user")
add_to_memory(bot_response, "assistant")

🚀 Usage

Start the Bot

python -m src.main

Test Memory System

python memory_test.py

Monitor Logs

The bot provides detailed logging for:

• Message processing
• Persona selection
• Memory operations
• AI service calls
• Error handling

📊 Configuration Options

Response Settings

'response_context_messages': 10,  # Number of messages for context
'max_tokens': 60,                 # AI response length limit
'persona_stickiness_duration': 180 # Persona consistency window (seconds)

Memory Settings

'memory_search_limit': 5,         # Number of relevant memories to retrieve
'memory_user_id': 'telegram_bot'  # Unique identifier for bot memory

🔧 Advanced Features

Persona Stickiness

• Maintains character consistency for 3 minutes
• 15% bonus to recently used personas
• Prevents rapid personality switching

Humanization Pipeline

• Raw AI responses → Grok fact gathering → OpenAI humanization
• Removes formal language and AI-speak
• Adds natural conversation patterns

Quote Removal

• Automatically removes surrounding quotes from responses
• Regex pattern: r'^"(.*)"$'
• Preserves internal quotations

Error Handling

• Graceful fallbacks for AI service failures
• Memory operation error recovery
• Comprehensive logging for debugging

🤝 Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests if applicable
5. Submit a pull request

📝 License

This project is licensed under the MIT License - see the LICENSE file for details.

🆘 Troubleshooting

Common Issues

Memory API Errors

# Check API key configuration
echo $MEM0_API_KEY

# Verify memory client initialization
python -c "from src.core_logic.memory import memory_client; print('Memory client OK')"

Persona Embeddings Missing

# Generate embeddings
python -c "from src.core_logic.llm_personas import PersonaManager; PersonaManager().generate_embeddings()"

Firestore Connection Issues

# Verify credentials
export GOOGLE_APPLICATION_CREDENTIALS=path/to/service-account.json

Debug Mode

Enable detailed logging by setting:

import logging
logging.basicConfig(level=logging.DEBUG)

📞 Support

For issues and questions:

1. Check the troubleshooting section
2. Review logs for error details
3. Create an issue with detailed information
4. Include relevant configuration (without API keys)

---

Note: This bot is designed for educational and experimental purposes. Ensure compliance with Telegram's Terms of Service and applicable regulations when deploying.