servicenow-mcp is a comprehensive MCP server for ServiceNow that lets AI agents connect to and orchestrate actions across ServiceNow modules. It exposes a wide toolset covering incidents, changes, scripting, ATF testing, flows, Now Assist integration, and multi-instance management, enabling your AI to read, create, update, deploy, and automate across multiple environments and customers. You can connect Claude, ChatGPT, Gemini, Cursor, Copilot, and other MCP-compatible clients to this server and issue plain-English requests that translate into ServiceNow actions. The server provides role-based tool packages, autonomous automation capabilities, and multi-tenant support so you can run workflows across dev, staging, and prod tenants from a single session.

To use the tools, configure the MCP connection with your ServiceNow instance credentials, then invoke the available MCP tools through your AI client. Examples include creating incidents, deploying server-side scripts, triggering flows, running ATF test suites, and performing cross-environment comparisons. Tools are organized by modules (e.g., Incident Management, Change Management, Scripting, ATF, and Reporting) and can be targeted to specific instances via instance switching when needed.

Prerequisites

• Node.js 20+ installed on your machine
• npm or pnpm installed
• Access to a ServiceNow instance with appropriate permissions for the tools you plan to use

1. Clone the repository git clone https://github.com/aartiq/servicenow-mcp.git cd servicenow-mcp

2. Install dependencies npm install

   or using pnpm

   pnpm install

3. Configure environment

   • Create a .env file or export the required variables listed below
   • Example (replace placeholders with real values): export SERVICENOW_INSTANCE="your-instance-name" export SERVICENOW_USERNAME="your-username" export SERVICENOW_PASSWORD="your-password" export MCP_LOG_LEVEL="info"

   Notes:

   • The server expects ServiceNow instance credentials to authenticate API calls. Use a dedicated service account if possible.
   • If you prefer not to use environment variables, you can supply them in the mcp_config under env.

4. Build the project (if needed) npm run build

   This will typically generate dist/server.js used by the MCP config

5. Start the MCP server npm start

   Or run the node command directly if you build output to dist/server.js

6. Verify the server is running

   • Check logs for startup messages indicating the MCP server is listening on the configured port
   • Use your MCP client to connect to the server-name (servicenow) and begin issuing commands

Tips and common issues:

• Ensure the ServiceNow instance URL and credentials have the necessary permissions for the actions you plan to perform (e.g., read access to incidents, ability to create/update records).
• If you encounter authentication errors, verify that the provided credentials are correct and that the target instance is reachable from the host running the MCP server.
• For multi-instance setups, you can switch instances per tool call or globally via the instance switching features supported by the server.
• Enable appropriate logging (MCP_LOG_LEVEL) during troubleshooting to capture detailed tool interactions.
• Keep your AI tool packages up to date to benefit from the latest capabilities and security improvements.
• Review documentation in docs/ for specifics on ATF, scripting, and Now Assist integrations to maximize tool usage.