Ghost Mesh Logo

A multi-agent chat system where conversations vanish, but digital souls persist. Built with FastAPI and Google's Gemini API.

• Multi-Agent Chat: Create and chat with multiple AI characters simultaneously
• Ephemeral Conversations: Chat sessions automatically expire after a configurable TTL
• Persistent AI Characters: Your AI characters persist across sessions
• Real-time WebSocket: Live chat experience with instant responses
• AI Avatar Generation: Generate unique avatars for your AI characters using Gemini's image generation
• Personality-driven AI: Each AI character has customizable personality traits
• Authentication: Secure JWT-based user authentication
• Scalable Architecture: Built with FastAPI, PostgreSQL, Redis, and Celery

• FastAPI: Modern, fast web framework for building APIs
• PostgreSQL: Primary database for user and AI character data
• Redis: Session management and message caching
• Celery: Background task processing for avatar generation
• Gemini API: Google's AI model for chat and image generation
• WebSocket: Real-time communication between clients and server

ghost_mesh/
├── ai_agent/ # AI character management and Gemini integration
├── auth/ # Authentication and authorization
├── chat/ # Chat session management and WebSocket handling
├── db/ # Database configuration and utilities
├── entities/ # SQLAlchemy models (User, AICharacter)
├── migrations/ # Alembic database migrations
├── tasks/ # Celery background tasks
├── users/ # User management
├── utils/ # Utility functions
├── app.py # FastAPI application entry point
├── config.py # Configuration settings
└── requirements.txt # Python dependencies

• Python 3.11+
• PostgreSQL database
• Redis server
• Google Gemini API key
• AWS S3 bucket (for image storage)

1. Clone the repository

   git clone <repository-url>
   cd ghost_mesh

2. Set up virtual environment

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

3. Install dependencies

   pip install -r requirements.txt

4. Configure environment variables

   cp .env.save .env
   # Edit .env with your actual configuration

5. Set up database

   # Run database migrations
   alembic upgrade head

Create a .env file with the following variables:

# Application
SECRET_KEY=your-secret-key
ALGORITHM=HS256
ACCESS_TOKEN_EXPIRE_MINUTES=220
# Database
DATABASE_URL=postgresql://user:password@host:port/database
# Gemini API
GEMINI_API_KEY=your-gemini-api-key
GEMINI_CHAT_MODEL=gemini-2.5-flash-lite
GEMINI_IMG_GEN_MODEL=imagen-4.0-generate-001
# Redis
REDIS_URL=redis://localhost:6379
REDIS_CELERY_BROKER=redis://localhost:6379
CHAT_SESSION_TTL=3600
# AWS S3
AWS_ENDPOINT_URL=your-s3-endpoint
AWS_ACCESS_KEY_ID=your-access-key
AWS_SECRET_ACCESS_KEY=your-secret-key
AWS_REGION_NAME=your-region
AWS_BUCKET_FOR_IMAGES=your-bucket-name
IMAGE_PUBLIC_ACCESS_URL=your-public-url

1. Start the API server

   python app.py

2. Start Celery worker (in another terminal)

   celery -A celery_app.celery_client worker --pool=threads -l info

1. Build and run with Docker Compose

   docker-compose up --build

• POST /auth/register - Register a new user
• POST /auth/login - User login and token generation

• POST /ai-character/create-character - Create a new AI character
• GET /ai-character/get-character/{id} - Get AI character details
• GET /ai-character/list-character - List user's AI characters
• DELETE /ai-character/delete-character - Delete an AI character
• PUT /ai-character/update-character - Update AI character

• POST /chat/create-session - Create a new chat session
• WebSocket /ws/chat/{session_id} - Real-time chat connection

• GET /users/profile - Get user profile
• PUT /users/update-profile - Update user profile

Users can create AI characters with:

• Name: Unique identifier for the character
• Description: Brief overview of the character
• Personality Traits: Detailed personality that influences AI responses
• Avatar: Auto-generated image based on character description

1. Session Creation: Users select AI characters to chat with
2. WebSocket Connection: Real-time bidirectional communication
3. Message Processing: Messages are stored in Redis with TTL
4. AI Responses: Each AI character responds based on their personality
5. Session Expiration: Sessions automatically expire after TTL

The system uses Gemini API with:

• System Instructions: Personality-based prompts
• Context Awareness: Full chat history for contextual responses
• Google Search: Enhanced responses with web search capabilities
• Concurrent Processing: Multiple AI characters respond simultaneously

Chat sessions automatically expire based on CHAT_SESSION_TTL (default: 3600 seconds). This ensures conversations remain ephemeral while AI characters persist.

AI characters follow this instruction pattern:

You are [character_name]. Mimic the following personality traits and answer the last message to the user only message no need of behavior, you can also chat with others characters. Personality trait: [personality_traits]. If you want to skip the conversation reply simply with a '.'

Run the test suite:

pytest tests/

The application includes Docker configuration for easy deployment:

• Dockerfile: Multi-stage build for production
• docker-compose.yml: Complete stack with API and Celery worker
• .dockerignore: Optimized container builds

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

This project is licensed under the MIT License.

• Voice chat integration
• AI character learning from conversations
• Multi-language support
• Advanced personality modeling
• Character relationship mapping
• Export/import character configurations
• Mobile app support

1. Database Connection: Ensure PostgreSQL is running and credentials are correct
2. Redis Connection: Verify Redis server is accessible
3. Gemini API: Check API key quota and model availability
4. WebSocket Issues: Verify CORS configuration and firewall settings

Check application logs for debugging:

# Development
python app.py
# Docker
docker-compose logs api
docker-compose logs celery

For support and questions:

• Create an issue on GitHub
• Check the troubleshooting section
• Review the API documentation

Ghost Mesh - Where digital souls linger and conversations fade away.