SafetyCulture MCP Server
Ask natural language questions about your SafetyCulture data using the SafetyCulture API.
SafetyCulture MCP Server
A Model Context Protocol (MCP) server for the SafetyCulture API. This project allows users to ask natural language questions about their SafetyCulture data after providing an API key.
Features
- Query SafetyCulture data using natural language
- Analyze inspection data and trends
- Compare safety metrics across time periods and categories
- Visualize inspection trends over time
Setup
- Clone this repository
- Install dependencies:
pip install -r requirements.txt - Copy
example.envto.envand configure your SafetyCulture API key - Run the server using one of these methods:
run_server.bat- Run the server with configuration from .env filerun_with_key.bat YOUR_API_KEY- Run the server with the provided API key
Testing the API
To test if your SafetyCulture API key works properly:
test_api.bat YOUR_API_KEY
Additional testing options:
test_api.bat- Run tests in interactive mode (prompts for API key)test_api.bat feed YOUR_API_KEY- Test just the Feed APItest_api.bat url- Check which API URLs are accessible without authentication
Usage with Claude for Desktop
- Install Claude for Desktop
- Configure Claude for Desktop to use this MCP server by editing the configuration file at
~/Library/Application Support/Claude/claude_desktop_config.json(Mac) or%APPDATA%\Claude\claude_desktop_config.json(Windows) - Add the following configuration:
{
"mcpServers": {
"safetyculture": {
"command": "python",
"args": [
"/path/to/your/project/src/main.py"
]
}
}
}
- Restart Claude for Desktop
- Use the MCP tools to query your SafetyCulture data with questions like:
- "How many inspections were done in this site over the last 3 months?"
- "Compare any trends in rise of injuries report for this category"
Available Tools
Authentication
authenticate: Authenticate with the SafetyCulture API using your API key
Inspection Data (Using Feed API)
get_inspections: Get SafetyCulture inspections for a specific time periodget_inspection_trends: Analyze trends in SafetyCulture inspections over timecompare_injury_reports: Compare injury reports between two time periods
Action Data (Using Feed API)
get_actions: Get SafetyCulture actions for a specific time period- Filter by status (e.g., 'in_progress', 'completed', 'overdue')
- Filter by priority (e.g., 'low', 'medium', 'high')
- View detailed information about each action
get_action_details: Get detailed information about a specific action by ID
About the Feed API
This MCP server uses the SafetyCulture Feed API, which provides a simple way to access collections of resources:
/feed/inspections: For listing inspections with various filter parameters/feed/actions: For listing actions with various filter parameters
The Feed API is preferred over the individual resource endpoints when you need to list multiple items.
Development
Project Structure
.
├── README.md
├── requirements.txt
├── example.env
└── src/
├── main.py # Main entry point
├── safetyculture_api/ # SafetyCulture API client
│ ├── __init__.py
│ └── client.py # API client implementation
├── tools/ # MCP tools
│ ├── __init__.py
│ └── inspection_tools.py # Inspection and action tools
└── utils/ # Utility modules
├── __init__.py
├── analysis.py # Data analysis utilities
├── config.py # Configuration management
└── date_utils.py # Date parsing utilities
Development Log
Initial Setup
- Created project structure
- Set up git repository
- Added README and requirements
- Implemented SafetyCulture API client
- Added MCP tools for querying inspection data
- Added utility modules for date parsing and data analysis
- Added configuration management
Related Servers
MCP Mempool
Access Bitcoin network data from mempool.space's API for AI agents and applications.
Skyvia MCP
Skyvia MCP Endpoint is a no-code gateway that makes your business data accessible to AI assistants in real time.
ServiceNow
An MCP server for interacting with ServiceNow instances to manage records, incidents, and other platform features.
Shopify MCP Server
Interact with Shopify store data using the GraphQL API.
WaveGuard
Physics-based anomaly detection via MCP — send any data, get anomalies back using wave-equation dynamics. No training pipelines, no model files.
Cloudflare MCP Server
A template for deploying a remote MCP server on Cloudflare Workers without authentication. It supports one-click deployment or local setup via npm.
Daraja MCP
Integrate AI applications with Safaricom's Daraja API for seamless interaction with M-Pesa services.
Zuora Product Catalog
An MCP server for accessing a Zuora product catalog, deployable on Cloudflare Workers.
Giant Swarm App Platform
A server for managing Giant Swarm App Platform deployments using Kubernetes credentials.
Weather MCP Server
Provides weather information using the free and open-source Open-Meteo API. No API key required.