# Diving Deep with Gemini: Exploring Intelligent Interactions through the Model Context Protocol

Welcome to an exciting exploration of how we can harness the power of Google's cutting-edge Gemini language models using the **Model Context Protocol (MCP)** framework. In this cookbook, we'll take a look under the hood at a project designed to facilitate intelligent interactions by exposing Gemini's capabilities as accessible tools within an MCP environment.

This cost-effective AI solution uses an MCP server-client architecture where affordable Gemini models act as clients and specialized models as servers, offering developers powerful yet affordable options for specific use cases.

## Authors

| Authors                                     |
| ------------------------------------------- |
| [KC Ayyagari](https://github.com/krishchyt) |
| [Para S](https://github.com/paraluke23)     |

## The Vision: Bridging the Gap with MCP

The goal of this project is to make interacting with complex language models like Gemini more structured and manageable. By leveraging the **Model Context Protocol (MCP)**, we can define specific functionalities of Gemini as distinct "tools." This allows a client application to intelligently decide when and how to utilize these powerful AI features based on the context of a conversation.

## Architecture

![mcp-gemini-architecture](https://storage.googleapis.com/github-repo/generative-ai/gemini/mcp/mcp-orchestration-app/mcp-gemini-architecture.svg)

## What You'll Need to Get Started

Before you can dive into this project, there are a few prerequisites you'll need to have in place:

- **Python Power:** You'll need Python 3.7 or a later version installed on your system.
- **Package Management with Pip:** Make sure you have pip, the Python package installer, ready to go.
- **Google Cloud Access:** This project relies on Gemini models. You'll need a Google Cloud Project with the Vertex AI API and Cloud Translation API enabled.
- **Authentication is Key:** Ensure you have the appropriate credentials configured for your Google Cloud Project. This could involve setting up environment variables or using a service account.

## Setting Up Your Environment

Ready to get your hands dirty? Here's a step-by-step guide to setting up your local environment:

1. **Clone the Code:** First things first, you'll need to grab the project code from its repository:

   ```bash
   git clone <repository-url>
   cd <repository-name>
   ```

2. **Set up `venv`:**

   ```bash
   python3 -m venv .venv
   source .venv/bin/activate
   ```

3. **Install the Magic Ingredients:** Next, let's install all the necessary Python libraries using pip:

   ```bash
   pip install -r requirements.txt
   ```

4. **Tell Us Your Secrets (Safely!):** We need to provide your Google Cloud Project details and potentially the specific Gemini model you want to use. Create a `.env` file in the root of the repository and add the following information, replacing the placeholders with your actual data:

   ```sh
   GOOGLE_CLOUD_PROJECT=your-google-cloud-project-id
   GOOGLE_CLOUD_LOCATION=your-google-cloud-region
   LLM_MODEL_NAME=gemma-3-27b-it
   GOOGLE_API_KEY="--Your Google AI Studio API Key for Gemma: https://aistudio.google.com/apikey --"
   ```

   **Important Note:** Make sure to add `.env` to your `.gitignore` file. You don't want to accidentally share your credentials!

5. **Reauthenticate gcloud if needed:**

   ```sh
   gcloud auth application-default login
   gcloud auth application-default set-quota-project <your-google-cloud-project-id>
   ```

6. **Enable Google Cloud APIs**

   Go to below URL(s) and enable them:

   - [Enable Google Translation API](https://console.developers.google.com/apis/api/translate.googleapis.com/overview)

7. **(Optional) Fine-Tune Your Server:** If you're using a `servers_config.json` file for server settings, ensure it's in the root directory and points to the `gemini_server.py` script correctly.

## Bringing the Application to Life

Now for the exciting part – running the application! This project has two main components: the MCP server and the client application.

### Starting the MCP Server

Open your terminal and navigate to the project directory. Then, execute the following command:

```bash
cd src
python gemini_client.py
```

![Output](https://storage.googleapis.com/github-repo/generative-ai/gemini/mcp/mcp-orchestration-app/Output.png)