kattastrofik-kattastrofey/my-new-repo

Meta-Agent Project

A flexible, modular AI agent platform that supports multiple LLM providers, including local deployment via Ollama.

Overview

Meta-Agent is a framework for creating, managing, and interacting with AI agents. It provides a consistent API interface regardless of which AI provider powers your agents, allowing you to:

• Create customized AI agents with specific personalities, knowledge, and capabilities
• Choose from multiple AI providers (Ollama, OpenAI, etc.)
• Manage users and their associated agents
• Save and reuse agent templates
• Interact with agents via text, with optional voice processing

Multi-Provider Architecture

Meta-Agent is designed with a provider-agnostic architecture that supports multiple AI backends:

Core Components:

1. Abstract AI Service Interface: Defines a consistent API for all AI providers
2. Provider Implementations:
   • Ollama Service: Integration with locally-hosted open-source models
   • OpenAI Service: Integration with OpenAI's commercial API
   • [Future Providers]: Easily extend with new providers (e.g., Gemini, Claude)
3. AI Service Factory: Creates the appropriate provider implementation based on configuration

Benefits:

• Provider Flexibility: Switch between providers without changing your application code
• Cost Control: Use free or local models by default, with commercial APIs as an option
• Future-Proofing: Easily add support for new providers as they emerge

Setup and Installation

Prerequisites

• Python 3.9+
• For local models: Ollama installed

Installation

1. Clone the repository:

   git clone <repository-url>
   cd meta-agent-project
   

2. Create and activate a virtual environment:

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

3. Install dependencies:

   pip install -r requirements.txt
   

4. Set up environment variables (or use defaults in config.py):

   # Create a .env file with your configuration
   touch .env
   

Configuration

The application uses a hierarchical configuration system with these options (in order of precedence):

1. Environment variables
2. .env file
3. Default values in config.py

Using Ollama (Default Provider)

Ollama is the default provider when no OpenAI API key is provided, offering free, locally-hosted models.

Setting Up Ollama

1. Install Ollama from ollama.ai
2. Pull a model (example):

   ollama pull llama2
   

3. Start Ollama server:

   ollama serve
   

Configuration for Ollama

In your .env file or environment variables:

OLLAMA_BASE_URL=http://localhost:11434  # Default Ollama server URL
OLLAMA_MODEL=llama2                     # Default model

Using OpenAI (Optional)

To use OpenAI as your provider:

1. Obtain an API key from OpenAI
2. Add to your .env file:

   OPENAI_API_KEY=your_api_key_here
   AI_PROVIDER=openai  # Override default provider
   

Project Structure

meta-agent-project/
├── main.py                  # Application entry point
├── README.md                # This documentation
├── requirements.txt         # Project dependencies
├── setup.sh                 # Setup script
├── src/
│   ├── api/                 # API endpoints and routes
│   │   ├── app.py           # FastAPI application
│   │   ├── auth/            # Authentication modules
│   │   └── routes/          # API routes (users, agents, templates)
│   ├── config.py            # Configuration management
│   ├── models/              # Data models and schemas
│   └── services/            # Service modules
│       ├── agent_customization/
│       ├── emotion_recognition/
│       ├── integration/     # External service integrations
│       │   ├── ai_service.py          # Abstract base class
│       │   ├── ai_service_factory.py  # Provider factory
│       │   ├── ollama_service.py      # Ollama implementation
│       │   └── openai_service.py      # OpenAI implementation
│       ├── storage/         # Database and persistence
│       │   └── database/
│       ├── template_management/
│       ├── user_management/
│       └── voice_processing/
└── tests/                   # Test modules

Basic Usage

Running the Application

python main.py

The API will be available at http://localhost:8000 by default.

API Endpoints

• Users: /api/users/ - User management
• Agents: /api/agents/ - Agent creation and management
• Templates: /api/templates/ - Template management
• Conversations: /api/conversations/ - Interact with agents

License

[License Information]

Contributing

[Contribution Guidelines]