API Now MCP Server

The API Now! Model Context Protocol (MCP) Server is a programmatic bridge for your AI tools (such as Claude Desktop, Cursor, or VS Code). It allows AI assistants to build, manage, and validate your backend data structures and endpoints directly from their chat interface.

Installation

You can install the MCP server package globally on your machine using npm:

npm install -g @api-now/mcp

Configuration & Authentication

The MCP server runs locally as a background process communicating over standard input/output (stdio).

Step 1: Authenticate via the CLI (Recommended)

Instead of managing API tokens manually in config files, the MCP server automatically reads your active CLI session. Authenticating via the CLI is the most practical setup:

Install the API Now! CLI globally:
Terminal window
```
npm install -g @api-now/cli
```
Log in to your account:
Terminal window
```
apinow auth login google
```
Find your Organization ID:
Terminal window
```
apinow orgs list
```
Set your default Organization:
Terminal window
```
apinow orgs set-default <organization_id>
```

Once set, the MCP server automatically inherits these credentials and targets your selected organization. To switch organizations at any time, simply run apinow orgs set-default again—there is no need to restart your editor or modify configuration files.

Editor Integration

1. Claude Desktop Setup

Add the server configuration to your Claude Desktop configuration file:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Linux: ~/.config/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json

Add the following config block (no environment variables are required since the server automatically reads your CLI profile):

{
  "mcpServers": {
    "api-now": {
      "command": "npx",
      "args": ["-y", "@api-now/mcp"]
    }
  }
}

2. Cursor Setup

To link the server with Cursor:

Open Cursor Settings, navigate to Features, and scroll to MCP.
Click + Add New MCP Server.
Configure the fields:
- Name: API Now
- Type: stdio
- Command: npx -y @api-now/mcp
- Note: Leave “Environment Variables” empty if you have completed the CLI login step.

Exposed Tool Directory

Once connected, your AI assistant will have access to the following toolsets:

Authentication Tools

configure_auth: Updates session keys, base API URL, and default organization.
get_auth_status: Checks active connection status and defaults.

Data Domain Tools

Schema Modifications: create_domain, read_domain, update_domain, delete_domain.
Database Table Mapping: add_entity, update_entity, remove_entity, add_property, update_property, remove_property.
Smart Fields: list_intelisense_helpers, apply_intelisense_helper.
Integrity Analysis: analyze_domain_deletion_impact (warns if removing a database column will break active APIs).

API Modeling Tools

Endpoint Configurations: create_api_model, read_api_model, update_api_model, delete_api_model.
Routes & Actions: expose_entity (projects database tables to web routes), add_entity_action (enables CRUD operations).
Schema Upgrades: analyze_api_domain_update (tests compatibility of database structure changes against exposed endpoints).

Catalog & Utility Tools

list_catalogs / search_public_catalog: Searches pre-made database schemas.
publish_catalog: Publishes stable schemas to the catalog.
generate_key: Pre-generates unique, platform-compliant database keys.