LLM Agent Framework MVP

A prototype implementation of an LLM agent-based framework using AutoGen for secure and contextual interactions between publisher and user agents.

Features

• User and Publisher agent interactions using AutoGen
• Mock article metadata and access control
• Query routing based on access levels
• Article matching and suggestion system

Setup

1. Create a virtual environment:

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

2. Install dependencies:

pip install -r requirements.txt

3. Run the MVP demo:

python main.py

Running Llama 3.2:3b Locally with Ollama

This project can utilize the Llama 3.2:3b model locally via Ollama. Follow these steps to set up Ollama and run the model:

1. Download and Install Ollama:

   • Visit the Ollama download page and install the application for your operating system (macOS, Windows, or Linux).

2. Start the Ollama Service:

   • Open a terminal and run:

     ollama serve

3. Download and Run the Llama 3.2:3b Model:

   • In a new terminal window, pull the model:

     ollama pull llama3.2:3b

   • To run the model interactively:

     ollama run llama3.2:3b

4. Integration:

   • Ensure Ollama is running and the model is available before starting the MVP demo if your workflow requires local LLM inference.

For more details, refer to the Ollama documentation.

Project Structure

/project-root/
├── agents/          # Agent definitions and configurations
├── data/           # Mock article metadata
├── logic/          # Core business logic
├── utils/          # Helper utilities
└── main.py        # MVP entry point

Development Status

MVP Implementation Phase - Basic agent interaction and content access control demonstration.