Django Firebase

Check out the project and install all Python dependencies including Django, Firebase Admin SDK, and LangChain packages.

git clone https://github.com/raghavdasila/django-firebase-mcp.git
cd django-firebase-mcp
pip install -r requirements.txt

In the Firebase Console, go to Project Settings > Service Accounts > Generate new private key. Download the JSON file and place it somewhere accessible on your server.

Create a .env file in the project root with your Firebase credentials path, storage bucket, and Django settings. Set MCP_TRANSPORT to 'stdio' for Claude Desktop or 'http' for networked agent access.

SERVICE_ACCOUNT_KEY_PATH=/path/to/your/firebase-service-account.json
FIREBASE_STORAGE_BUCKET=your-project-id.appspot.com
MCP_TRANSPORT=stdio
MCP_HOST=0.0.0.0
MCP_PORT=8000
DEBUG=True
SECRET_KEY=your-django-secret-key-here
USE_REDIS=false
# REDIS_URL=redis://localhost:6379/0

Apply database migrations and then start the Django development server which also serves the MCP endpoint.

python manage.py migrate
python manage.py runserver

Run the built-in standalone agent script to verify your Firebase credentials and MCP tools are working correctly before connecting an AI client.

python firebase_admin_mcp/standalone_firebase_agent.py

Add the server to your Claude Desktop config to run it as a local stdio process. Claude will spawn the Django server automatically.

{
  "mcpServers": {
    "django-firebase": {
      "command": "python",
      "args": ["manage.py", "runserver"],
      "env": {
        "SERVICE_ACCOUNT_KEY_PATH": "/path/to/firebase-service-account.json",
        "FIREBASE_STORAGE_BUCKET": "your-project-id.appspot.com",
        "MCP_TRANSPORT": "stdio"
      }
    }
  }
}