# Claude-Gemini Collaborative Integration

This project enables seamless collaboration between Claude Code and Gemini CLI, allowing them to work together on tasks, discuss ideas, and refine solutions in real-time. This integration leverages the Model Context Protocol (MCP) to facilitate communication and tool usage between the two AI agents.

## Features

*   **Start Collaborative Sessions**: Initiate a new discussion with Gemini on a specific topic.
*   **Consult Gemini**: Ask questions or provide context to Gemini within an ongoing collaboration.
*   **Retrieve Conversation History**: Access the full transcript of a collaborative session.
*   **Context Preservation**: Gemini remembers previous interactions within a session, allowing for iterative discussions.
*   **Rate Limiting Handling**: The collaborative server automatically manages rate limits for Gemini API calls.

## Prerequisites

Before you begin, ensure you have the following installed:

*   **Python 3.8+**: The project is built with Python.
*   **`pip`**: Python package installer.
*   **Claude Desktop**: The application that hosts Claude Code.
*   **Gemini CLI**: Ensure the Gemini CLI is installed and configured with your API key. You can verify its installation by running `gemini --version` in your terminal.

## Setup Guide

Follow these steps to set up the Claude-Gemini collaborative integration:

### 1. Clone the Repository

First, clone this repository to your local machine:

```bash
git clone <repository_url>
cd claude-gemini-integration
```
*(Replace `<repository_url>` with the actual URL of your repository)*

### 2. Create and Activate a Python Virtual Environment

It's recommended to use a virtual environment to manage dependencies:

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

### 3. Install Dependencies

Install the required Python packages using `pip`:

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

### 4. Start the Collaborative Server

The collaborative server acts as an intermediary between Claude Code and Gemini. It needs to be running in the background.

```bash
# Navigate to the project directory if you're not already there
cd /Users/jamiearonson/Documents/claude-gemini-integration

# Activate your virtual environment
source venv/bin/activate

# Start the server in the background
python tools/mcp/collaborative-server.py &
```
You can verify the server is running by checking its health endpoint:
```bash
curl http://localhost:8080/health
```
You should see a response like `{"status": "ok"}`.

### 5. Configure Claude Desktop for MCP

You need to tell Claude Desktop about the new MCP server. This involves adding a configuration snippet to your `claude_desktop_config.json` file.

**Locate your `claude_desktop_config.json` file:**
*   **macOS**: `~/Library/Application Support/Claude/claude_desktop_config.json`
*   **Windows**: `%APPDATA%\Claude\claude_desktop_config.json`
*   **Linux**: `~/.config/Claude/claude_desktop_config.json`

**Add the following JSON snippet** to the `mcpServers` section of your `claude_desktop_config.json` file. If the `mcpServers` section doesn't exist, you can create it.

```json
{
  "mcpServers": {
    "gemini-collaboration": {
      "command": "python3",
      "args": ["/Users/jamiearonson/Documents/claude-gemini-integration/mcp_server.py"],
      "env": {
        "MCP_PORT": "8080"
      }
    }
  }
}
```
**Important**: Ensure the `args` path points to the absolute path of `mcp_server.py` on your system.

### 6. Restart Claude Desktop

For the configuration changes to take effect, you must close and reopen the Claude Desktop application.

### 7. Verify MCP Tools are Available

After restarting Claude Desktop, in a new Claude Code session, you should see these tools available:

*   `start_gemini_collaboration`: Start a new collaborative conversation with Gemini.
*   `consult_gemini`: Ask Gemini a question in the current collaboration context.
*   `get_collaboration_history`: Get the full conversation history with Gemini.

## Usage

Once set up, you can interact with Gemini directly through Claude Code using natural language.

### Starting a Collaboration

To begin a new collaborative session with Gemini, simply tell Claude Code:

```
"Start a collaboration with Gemini about designing a REST API."
```
Claude Code will automatically use the `start_gemini_collaboration` tool.

### Having a Discussion

Once a collaboration is active, you can ask Gemini questions or provide further context:

```
"Ask Gemini what they think about using microservices vs monolith for this project."
```
Claude Code will use the `consult_gemini` tool to get Gemini's input, maintaining the conversation context.

### Getting History

To review the entire conversation history of the current collaboration:

```
"Show me the collaboration history with Gemini."
```
This will use the `get_collaboration_history` tool.

### Example Workflow

Here's a typical interaction flow:

1.  **You**: "I need to design a database schema for an e-commerce site. Start a collaboration with Gemini."
2.  **Claude**: (Uses `start_gemini_collaboration` tool, confirms collaboration started)
3.  **You**: "Ask Gemini about the best approach for handling product variants (e.g., size, color)."
4.  **Claude**: (Uses `consult_gemini` tool, displays Gemini's response)
5.  **You**: "What does Gemini think about our payment table design, specifically regarding PCI compliance?"
6.  **Claude**: (Continues the collaboration, building on previous context, and provides Gemini's insights)
7.  **You**: "Show me the full collaboration history."
8.  **Claude**: (Displays the entire conversation transcript)

## Troubleshooting

### Tools Not Available in Claude Code

*   Ensure Claude Desktop was fully restarted after modifying `claude_desktop_config.json`.
*   Verify that the `claude_desktop_config.json` file exists at the correct path for your operating system and that the JSON is valid.
*   Double-check that the `command` and `args` paths in `claude_desktop_config.json` are absolute and correct.

### Collaborative Server Connection Issues

*   **Check if the server is running**:
    ```bash
    curl http://localhost:8080/health
    ```
    If it's not running, restart it as described in [Step 4](#4-start-the-collaborative-server).
*   **Check for port conflicts**: Ensure no other application is using port `8080`.

### Gemini CLI Issues

*   **Verify Gemini CLI installation**: Run `gemini --version` in your terminal.
*   **Check API key configuration**: Ensure your Gemini API key is correctly set up and accessible by the Gemini CLI.

## Testing

You can run the provided test script to verify the MCP client and server functionality:

```bash
# Navigate to the project directory
cd /Users/jamiearonson/Documents/claude-gemini-integration

# Activate your virtual environment
source venv/bin/activate

# Run the tests
python test_mcp.py
```

## Benefits of Collaboration

This integration provides significant benefits for complex software engineering tasks:

*   **Real-time Discussion**: Engage in actual back-and-forth conversations with Gemini, not just one-off queries.
*   **Contextual Understanding**: Gemini maintains context throughout the conversation, leading to more relevant and insightful responses.
*   **Iterative Problem Solving**: Work through problems step-by-step, refining ideas and solutions collaboratively.
*   **Enhanced Problem Solving**: Leverage the strengths of both Claude and Gemini to tackle challenging problems more effectively.
*   **Seamless Workflow**: Integrate collaborative AI assistance directly into your Claude Code environment.