ShotGrid

shotgrid-mcp-rs provides a modern Rust-based MCP server that lets large language models interact with ShotGrid through a dedicated MCP interface. It supports a comprehensive set of CRUD operations, search, time tracking, and advanced aggregation, while offering automatic input validation and sanitization to reduce common API errors. This server can be run locally in stdio mode for desktop clients or exposed via HTTP stream for remote clients, enabling efficient integration with LLM workflows.

How to use

You run the MCP server locally in stdio mode for desktop clients, or switch to HTTP stream mode to expose the MCP endpoints to remote clients. In stdio mode, launch the server and provide the necessary ShotGrid credentials via environment variables. In HTTP stream mode, you enable the HTTP endpoint and connect from remote clients using the /mcp channel.

How to install

Prerequisites: install Rust toolchain and Cargo, since you will build or install the MCP server from source or from crates.io.

# Prerequisites: install Rust
# If you already have Rust, skip this step
rustup default stable

# Option 1: Install from crates.io (recommended)
cargo install shotgrid-mcp-rs --all-features

# Option 2: Build from source
# 1) Clone the repository
git clone https://github.com/ssoj13/shotgrid-mcp-rs
cd shotgrid-mcp-rs

# 2) Set up environment variables for ORM code generation (if you plan to use ORM features)
export SG_SERVER="https://yoursite.shotgrid.autodesk.com"
export SG_SCRIPT_NAME="your_script_name"
export SG_SCRIPT_KEY="your_script_key"

# 3) Generate ORM models and install
./bootstrap.ps1 codegen
cargo install --path . --all-features

# After installation, the binary will be available in your Cargo bin directory (e.g., ~/.cargo/bin/shotgrid-mcp-rs)

Configuration and usage notes

Two transport modes are supported. In stdio mode you run the binary directly and connect with local clients. In HTTP stream mode you start the server with streaming enabled and connect remotely.

{
  "mcpServers": {
    "shotgrid": {
      "type": "stdio",
      "name": "shotgrid",
      "command": "shotgrid-mcp-rs",
      "args": [],
      "env": {
        "SG_SERVER": "https://yoursite.shotgrid.autodesk.com",
        "SG_SCRIPT_NAME": "your_script_name",
        "SG_SCRIPT_KEY": "your_script_key"
      }
    }
  }
}

Using the MCP server with Claude Desktop

Configure Claude Desktop to use the local MCP server in stdio mode. Provide the ShotGrid connection details as environment variables in the MCP server entry.

{
  "mcpServers": {
    "shotgrid": {
      "command": "shotgrid-mcp-rs",
      "env": {
        "SG_SERVER": "https://yoursite.shotgrid.autodesk.com",
        "SG_SCRIPT_NAME": "your_script_name",
        "SG_SCRIPT_KEY": "your_script_key"
      }
    }
  }
}

MCP Tools overview

The MCP server exposes a suite of 14 tools to manage ShotGrid data and workflows. This includes the core CRUD operations, time logging, cross-entity search, aggregations, and note threads.

Security and best practices

Use strong, unique script keys and restrict access to the MCP endpoint. Rotate credentials periodically and prefer the stdio mode for local development. If you enable HTTP streaming for remote access, ensure you operate behind appropriate network security measures and monitor logs for unusual activity.

Troubleshooting and notes

If you encounter API field errors, ensure input sanitization is active so field names are trimmed of leading or trailing whitespace. This improves compatibility with various MCP clients and ShotGrid endpoints.

Available tools

schema_read

Get entity field schemas

search_entities

Search with filters and pagination across entities

find_one

Find a single entity by criteria

read_entity

Read an entity by ID with optional fields

create_entity

Create a new entity with specified fields

update_entity

Update an existing entity by ID with new data

delete_entity

Soft delete an entity by ID

revive_entity

Restore a previously deleted entity

read_timelogs

Query time log entries and related data

timelog_stats

Aggregate time log statistics

text_search

Cross-entity text search across multiple types

summarize

Aggregate data with grouping and calculations

batch

Perform bulk operations in a single request

note_thread_read

Retrieve note threads with replies and attachments