Overview
What is ServiceNow MCP Server?
The ServiceNow MCP Server is a Model Context Protocol integration that enables AI models to interact directly with ServiceNow instances. It provides a set of tools for automating incident management, knowledge base article creation, record producer setup, and other common ServiceNow operations. This server is intended for developers building AI-powered workflows that require read/write access to ServiceNow data.
How to use ServiceNow MCP Server?
Configure environment variables (SERVICENOW_INSTANCE, SERVICENOW_USERNAME, SERVICENOW_PASSWORD, optional GEMINI_API_KEY) in a .env file, install dependencies with pip install -e ., then start the server using mcp install <file_name>.py. The server exposes MCP tools such as create_incident, create_kb_article, and create_record_producer.
Key features of ServiceNow MCP Server
- Direct integration with any ServiceNow instance
- Supports creating incidents, KB articles, client scripts, business rules, SLA definitions, record producers, and variable sets
- Environment‑based secure credential management
- Built on FastAPI for high‑performance async operations
- Optional Gemini API key for AI capabilities
Use cases of ServiceNow MCP Server
- Automatically create ServiceNow incidents from AI‑powered chat or alerts
- Generate knowledge base articles through natural language prompts
- Set up record producers with custom form fields and server‑side scripting
- Create reusable variable sets for catalog items across multiple services
- Let AI agents manage business rules, SLA definitions, and client scripts
FAQ from ServiceNow MCP Server
What distinguishes this server from direct ServiceNow API calls?
This server wraps ServiceNow REST API operations as MCP tools, making them natively available to any MCP‑compatible AI client. It handles authentication and error handling internally so AI models can focus on the task.
What runtime and dependencies are required?
Python 3.12 or higher, a ServiceNow instance with API access, valid credentials, and the FastAPI and uvicorn libraries.
Where does the server store credentials and data?
Credentials are read from a .env file (never hardcoded). All operational data remains in your ServiceNow instance; the server uses the instance’s REST API to create and read data.
How does authentication and transport work?
Authentication uses username/password from environment variables. The server runs as a FastAPI application over HTTP (typically stdio for MCP clients, but FastAPI enables standard HTTP transport if needed). An optional Gemini API key is used only for AI features.
Can I add new ServiceNow operations?
Yes. Add a new async function in servicenow.py, decorate it with @mcp.tool(), include proper documentation and error handling, then test the new tool.
