Britive's MCP Server enables AI agents and users to interact with the Britive platform. Britive's MCP server exposes several tools that enable users and AI agents to interact with the Britive platform for dynamic access, query configurations, reporting, and access activity.

To learn more about MCPs, see [Get Started with MCP](https://modelcontextprotocol.io/docs/getting-started/intro).

---

## Prerequisites

- Python version 3.10 or higher. [Python downloads](https://www.python.org/downloads/).
- Ensure you have the latest version of Git installed and access to the repository. Download git from the [Git downloads](https://git-scm.com/downloads).
- Any MCP client. For example: Claude desktop, VS Code Copilot

---

## Setting up the Britive MCP Server

1. Clone the GitHub repository that has the MCP server.

```shell
git clone https://github.com/britive/mcp-server.git
````

2. Change directory to `mcp-server`

```shell
cd mcp-server
```

3. Create a virtual environment

```shell
python -m venv <virtual_env_name>
```

4. Activate the virtual environment

* **Windows**

  ```shell
  <virtual_env_name>\Scripts\activate
  ```
* **Linux**

  ```shell
  source <virtual_env_name>/bin/activate
  ```

5. Install Python packages (dependencies)

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

---

## Configuring your MCP client and Authentication:
Authentication to the Britive platform can be performed using the Britive CLI or a Static Token

* #### **Option 1: CLI login using PyBritive (Recommended):**

1. Install the PyBritive CLI if not installed. For more information, see [PyBritive Documentation](https://britive.github.io/python-cli/)

2. Configure the tenant, if you have not configured it already, using the following
command. To find out more about how to enter the tenant name, see [PyBritive: Tenant selection logic](https://britive.github.io/python-cli/#tenant-selection-logic):

   ```shell
   pybritive configure tenant
   ```

3. Log in to PyBritive:

   ```shell
   pybritive login
   ```

4. Log in using the following command if you have multiple tenants:

   ```shell
   pybritive login --tenant=<your_tenant_name>
   ```
5. Log in directs you to the Britive login page, where you can enter your credentials. 
6. Modify your MCP JSON file to configure the Britive MCP.
> **Note:**  
> For Windows, use **backslash (\\\\)**.  
> For macOS/Linux, use a **forward slash (/)** in the path.
```json
{
  "servers": {
    "britive": {
      "command": "C:\\Users\\YourName\\mcp-server\\venv\\Scripts\\python.exe",
      "args": [
        "C:\\Users\\YourName\\mcp-server\\britive_mcp_tools\\core\\mcp_runner.py"
      ],
      "env": {
        "PYTHONPATH": "C:\\Users\\YourName\\mcp-server",
        "BRITIVE_TENANT": "your_tenant_name"
      }
    }
  }
}
```

Where:
<table>
  <tbody>
    <tr>
      <td><code>command</code></td>
      <td>Full path to the Python executable</td>
    </tr>
    <tr>
      <td><code>args</code></td>
      <td>Full path to the <code>britive_mcp_tools\core\mcp_runner.py</code> module.</td>
    </tr>
    <tr>
      <td><code>PYTHONPATH</code></td>
      <td>Set the Python path to the <code>mcp-server</code> directory (repository)</td>
    </tr>
    <tr>
      <td><code>BRITIVE_TENANT</code></td>
      <td>Your tenant on Britive. This tenant name must match the name used in the PyBritive CLI</td>
    </tr>
  </tbody>
</table>

* #### **Option 2: Login using a static token:**

1. Generate a static token from Britive UI. For more information, see [API Tokens](https://docs.britive.com/v1/docs/api-tokens-1).
2. Copy the token to use in the MCP JSON file.
3. Modify your MCP JSON file to configure the Britive MCP.

> **Note:**  
> For Windows, use **backslash (\\\\)**.  
> For macOS, use a **forward slash (/)** in the path.
```json
{
  "servers": {
    "britive": {
      "command": "C:\\Users\\YourName\\mcp-server\\venv\\Scripts\\python.exe",
      "args": [
        "C:\\Users\\YourName\\mcp-server\\britive_mcp_tools\\core\\mcp_runner.py"
      ],
      "env": {
        "PYTHONPATH": "C:\\Users\\YourName\\mcp-server",
        "BRITIVE_TENANT": "your_tenant_name",
        "BRITIVE_STATIC_TOKEN": "your_static_token_here"
      }
    }
  }
}
```

**Where:**
<table>
  <tbody>
    <tr>
      <td><code>command</code></td>
      <td>Full path to the Python executable</td>
    </tr>
    <tr>
      <td><code>args</code></td>
      <td>Full path to the <code>britive_mcp_tools\core\mcp_runner.py</code> module.</td>
    </tr>
    <tr>
      <td><code>PYTHONPATH</code></td>
      <td>Set the Python path to the <code>mcp-server</code> directory (repository)</td>
    </tr>
    <tr>
      <td><code>BRITIVE_TENANT</code></td>
      <td>Your tenant on Britive. Do not use the entire FQDN for the subdomain. For example, for 
      <code>https://super-customer.test.aws.britive-corp.com</code>, use 
      <code>super-customer.test.aws</code> as a tenant subdomain.
      </td>
    </tr>
    <tr>
      <td><code>BRITIVE_STATIC_TOKEN</code></td>
      <td>Static token created in the previous step</td>
    </tr>
  </tbody>
</table>

 
## Connect to the MCP server using a client application. 
For more information, see [Connect to Local MCP Servers](https://modelcontextprotocol.io/quickstart/user).


## Using On-Behalf-Of MCP Functionality
To use OBO MCP, you can following environment variable: 

<table>
  <tbody>
    <tr>
      <td><code>BRITIVE_EMAIL</code></td>
      <td>The Britive email of the person you want the MCP server to impersonate</td>
    </tr>
  </tbody>
</table>

If this environment variable is set, the MCP server will assume the user intends to use OBO and will take priority over non OBO MCP functionality. The `BRITIVE_STATIC_TOKEN` should refer to the following service identity that the user would like to perform impersonation. Currently, not all tools are supported by OBO.

List of supported OBO tools:
1. `my_access`