Agentic Gemini

Agentic Gemini is an exploratory repository demonstrating multi-agent workflows using the AG2 (autogen) framework and Google's Gemini models (via AI Studio).

The primary objective is to provide a hands-on, runnable example of various agent patterns, from simple code execution and group chats to complex, human-in-the-loop validation, all running securely within a Docker environment.

Table of Contents

• Agentic Gemini

• Table of Contents

• Getting Started

  • Prerequisites
  • Configuration
  • Running the Application (Docker)
  • Local Development Setup (Optional)

• Acknowledgement

Getting Started

Follow these instructions to get a copy of the project up and running on your local machine.

Prerequisites

You will need the following tools installed on your system:

• Git
• Python 3.11+
• Docker Engine

Configuration

The application requires a Google Gemini API key. This project uses a .gitignore to protect your keys, so you must create two files:

1. config.json: Create this file in the root directory, using sample_config.json as a template. This file will contain your actual API key.

   [
     {
       "model": "gemini-2.5-flash-lite",
       "api_key": "your-AI-Studio-API-key-goes-here",
       "api_type": "google"
     }
   ]

2. config_path.json: Create this file in the root directory, using sample_config_path.json as a template. This file tells the application where to find your config.json. For this project, you can just set it to config.json.

   {
     "config_path": "config.json"
   }

Running the Application (Docker)

This is the recommended way to run the application. It ensures a consistent, secure, and isolated environment.

1. Clone the repository:

   git clone [https://github.com/your-username/Agentic-Gemini.git](https://github.com/your-username/Agentic-Gemini.git)
   cd Agentic-Gemini

2. Build the Docker image: This command reads the Dockerfile and builds a local image named agentic-gemini.

   docker build -t agentic-gemini .

3. Run the application: This command runs the app interactively.

   • The -v /var/run/docker.sock:/var/run/docker.sock flag is required; it allows the agent inside the container to spawn new Docker containers for code execution.

   • The -v "/your/target/directory":"/my_files" flag is required for Mode 5. This mounts your local directory (e.g., ~/Documents) into the container at /my_files, allowing the agent to find, read, edit, and run files.

     • Note: This flag must be read-write (without the :ro suffix). The agent needs write-access to this directory to create the temporary scripts it uses for code execution, even if you do not plan to edit your original files.

   docker run -it --rm -v /var/run/docker.sock:/var/run/docker.sock -v "/your/local/directory":"/my_files" agentic-gemini

Local Development Setup (Optional)

If you want to edit the code locally with IDE support (like linting and code completion in VS Code), you should set up a local virtual environment.

1. Navigate to the project directory:

   cd Agentic-Gemini

2. Create a virtual environment:

   python -m venv venv

3. Activate the environment:

   • Linux/macOS:

     source venv/bin/activate

   • Windows (PowerShell):

     .\venv\Scripts\Activate.ps1

4. Install dependencies:

   pip install -r requirements.txt

   You can now open the folder in your IDE. Remember to run the application using the Docker commands above.

Acknowledgement

This project is built upon the AG2 (autogen) framework.

• ag2ai/ag2 Repository