Client Setup Guide
Step-by-step setup for connecting NowAIKit to each supported AI client. Choose your client below and follow the configuration instructions to get up and running in minutes.
Prerequisites
All clients require the following before you begin:
Check with node --version. Download from nodejs.org.
Run npm install && npm run build in your NowAIKit directory.
Developer, sandbox, or production instance. You need the full URL and a service account.
Claude Code
Claude Code discovers MCP servers via the claude mcp add command. Run one of the commands below in your terminal to register NowAIKit.
Basic Auth
$ claude mcp add servicenow \
--command "node /absolute/path/to/nowaikit/dist/server.js" \
--env SERVICENOW_INSTANCE_URL=https://yourinstance.service-now.com \
--env SERVICENOW_AUTH_METHOD=basic \
--env SERVICENOW_BASIC_USERNAME=your_username \
--env SERVICENOW_BASIC_PASSWORD=your_password \
--env WRITE_ENABLED=false
OAuth
$ claude mcp add servicenow \
--command "node /absolute/path/to/nowaikit/dist/server.js" \
--env SERVICENOW_INSTANCE_URL=https://yourinstance.service-now.com \
--env SERVICENOW_AUTH_METHOD=oauth \
--env SERVICENOW_OAUTH_CLIENT_ID=your_client_id \
--env SERVICENOW_OAUTH_CLIENT_SECRET=your_client_secret \
--env SERVICENOW_OAUTH_USERNAME=your_username \
--env SERVICENOW_OAUTH_PASSWORD=your_password \
--env WRITE_ENABLED=false
/absolute/path/to/nowaikit/dist/server.js with the actual absolute path on your system. Run pwd in the NowAIKit directory to find it.
Claude Desktop
Edit your Claude Desktop configuration file to add NowAIKit as an MCP server. The file location depends on your operating system.
Config file locations
| OS | Path |
|---|---|
| macOS | ~/Library/Application Support/Claude/claude_desktop_config.json |
| Windows | %APPDATA%\Claude\claude_desktop_config.json |
| Linux | ~/.config/Claude/claude_desktop_config.json |
Basic Auth Config
{
"mcpServers": {
"servicenow": {
"command": "node",
"args": ["/absolute/path/to/nowaikit/dist/server.js"],
"env": {
"SERVICENOW_INSTANCE_URL": "https://yourinstance.service-now.com",
"SERVICENOW_AUTH_METHOD": "basic",
"SERVICENOW_BASIC_USERNAME": "your_username",
"SERVICENOW_BASIC_PASSWORD": "your_password",
"WRITE_ENABLED": "false",
"MCP_TOOL_PACKAGE": "service_desk"
}
}
}
}
OAuth Config
{
"mcpServers": {
"servicenow": {
"command": "node",
"args": ["/absolute/path/to/nowaikit/dist/server.js"],
"env": {
"SERVICENOW_INSTANCE_URL": "https://yourinstance.service-now.com",
"SERVICENOW_AUTH_METHOD": "oauth",
"SERVICENOW_OAUTH_CLIENT_ID": "your_client_id",
"SERVICENOW_OAUTH_CLIENT_SECRET": "your_client_secret",
"SERVICENOW_OAUTH_USERNAME": "your_username",
"SERVICENOW_OAUTH_PASSWORD": "your_password",
"WRITE_ENABLED": "false"
}
}
}
}
OpenAI Codex
The OpenAI integration uses a Python wrapper that spawns the MCP server and translates tool schemas to OpenAI function definitions. This supports all function-calling models including GPT-4.1, o3, and o4-mini.
Setup
$ cd clients/codex
$ pip install openai python-dotenv
Run
$ python servicenow_openai_client.py
.env file contains your OPENAI_API_KEY and ServiceNow credentials before running the client.
Google Gemini / Vertex AI
Connect NowAIKit to Google Gemini and Vertex AI. The integration converts tool schemas to Gemini function declarations, supporting Gemini 2.5 Pro, Gemini 2.5 Flash, and all Gemini models with function calling.
Gemini API Setup
$ cd clients/gemini
$ pip install google-generativeai python-dotenv
$ python servicenow_gemini_client.py
Vertex AI Setup
For enterprise Vertex AI deployments, authenticate with Google Cloud first:
$ gcloud auth application-default login
Cursor
Create or edit .cursor/mcp.json in your project root to register NowAIKit:
{
"mcpServers": {
"servicenow": {
"command": "node",
"args": ["/absolute/path/to/nowaikit/dist/server.js"],
"env": {
"SERVICENOW_INSTANCE_URL": "https://yourinstance.service-now.com",
"SERVICENOW_AUTH_METHOD": "basic",
"SERVICENOW_BASIC_USERNAME": "your_username",
"SERVICENOW_BASIC_PASSWORD": "your_password",
"WRITE_ENABLED": "false"
}
}
}
}
Alternatively, add it globally via Cursor Settings → MCP → Add new global server.
VS Code
VS Code MCP support requires the GitHub Copilot extension or Claude for VS Code (Cline). Create or edit .vscode/mcp.json in your project root:
{
"servers": {
"servicenow": {
"type": "stdio",
"command": "node",
"args": ["/absolute/path/to/nowaikit/dist/server.js"],
"env": {
"SERVICENOW_INSTANCE_URL": "https://yourinstance.service-now.com",
"SERVICENOW_AUTH_METHOD": "basic",
"SERVICENOW_BASIC_USERNAME": "your_username",
"SERVICENOW_BASIC_PASSWORD": "your_password",
"WRITE_ENABLED": "false"
}
}
}
}
"servers" key (not "mcpServers") and requires the "type": "stdio" property for each server entry.
Troubleshooting
Common issues and their solutions when setting up AI clients with NowAIKit:
| Problem | Solution |
|---|---|
| Server won't start | Run npm run build first; check Node.js version is 20+ with node --version |
| Authentication failed | Verify instance URL has no trailing slash; confirm credentials are correct |
| No tools showing | Check MCP server logs; verify the path to server.js is absolute, not relative |
| Tool call errors | Check WRITE_ENABLED, SCRIPTING_ENABLED, and ATF_ENABLED flags match your intended permissions |
| OAuth token errors | Verify client ID and client secret; ensure the OAuth application is active in ServiceNow |
.env file or config files containing credentials. Add them to .gitignore and use a dedicated service account with the minimum required ServiceNow roles.



