Quickstart Guide

Get started with ShotGrid MCP Server in minutes

Quickstart Guide

This guide will help you set up a basic ShotGrid MCP Server and create your first tools. By the end, you’ll have a working server that can interact with ShotGrid data through the MCP protocol.

Basic Server Setup

First, let’s create a simple server that connects to ShotGrid:

from shotgrid_mcp_server import ShotGridMCPServer

# Create a server with your ShotGrid credentials
server = ShotGridMCPServer(
    name="ShotGrid Assistant",
    shotgrid_url="https://your-site.shotgunstudio.com",
    script_name="your_script_name",
    api_key="your_api_key"
)

# Run the server
if __name__ == "__main__":
    server.run(host="localhost", port=8000)

Save this as server.py and run it with python server.py. Your server will start on http://localhost:8000.

Using Mockgun for Testing

For development and testing, you can use Mockgun instead of connecting to a real ShotGrid instance:

from shotgrid_mcp_server import ShotGridMCPServer

# Create a server with Mockgun
server = ShotGridMCPServer(
    name="ShotGrid Test Server",
    use_mockgun=True,  # This enables Mockgun
    schema_path="path/to/schema.bin"  # Optional: path to a schema file
)

if __name__ == "__main__":
    server.run(host="localhost", port=8000)

Creating Your First Tool

Let’s add a tool to search for projects:

from shotgrid_mcp_server import ShotGridMCPServer

server = ShotGridMCPServer(
    name="ShotGrid Assistant",
    use_mockgun=True  # For testing
)

@server.tool()
def find_projects(status: str = None):
    """
    Find projects in ShotGrid, optionally filtered by status.
    
    Args:
        status: Filter projects by status (e.g., "Active", "Archived")
    
    Returns:
        A list of projects matching the criteria
    """
    filters = []
    if status:
        filters.append(["sg_status", "is", status])
    
    return server.connection.find(
        "Project",
        filters,
        ["id", "name", "code", "sg_status"]
    )

if __name__ == "__main__":
    server.run(host="localhost", port=8000)

Adding Test Data to Mockgun

If you’re using Mockgun, you’ll need to add some test data:

from shotgrid_mcp_server import ShotGridMCPServer

server = ShotGridMCPServer(
    name="ShotGrid Test Server",
    use_mockgun=True
)

# Add test data to Mockgun
@server.on_startup
def create_test_data():
    # Create test projects
    server.connection.create("Project", {
        "name": "Awesome Film",
        "code": "AWSM",
        "sg_status": "Active"
    })
    
    server.connection.create("Project", {
        "name": "Old Project",
        "code": "OLD",
        "sg_status": "Archived"
    })

@server.tool()
def find_projects(status: str = None):
    """Find projects in ShotGrid, optionally filtered by status."""
    filters = []
    if status:
        filters.append(["sg_status", "is", status])
    
    return server.connection.find(
        "Project",
        filters,
        ["id", "name", "code", "sg_status"]
    )

if __name__ == "__main__":
    server.run(host="localhost", port=8000)

Testing Your Server

You can test your server using the built-in MCP client:

from mcp.client import Client

async def test_server():
    # Connect to your server
    client = Client("http://localhost:8000")
    
    # List available tools
    tools = await client.list_tools()
    print(f"Available tools: {[tool.name for tool in tools]}")
    
    # Call the find_projects tool
    result = await client.call_tool("find_projects", {"status": "Active"})
    print(f"Active projects: {result}")

if __name__ == "__main__":
    import asyncio
    asyncio.run(test_server())

Save this as test_client.py and run it with python test_client.py while your server is running.

Next Steps

Now that you have a basic server running, you can:

• Add more tools for different ShotGrid entities
• Learn about optimized queries for better performance
• Explore batch operations for efficient data manipulation
• Set up connection pooling for production use

Congratulations! You’ve created your first ShotGrid MCP Server. Continue exploring the documentation to learn more about the server’s capabilities.

Assistant

Responses are generated using AI and may contain mistakes.