NetSuite MCP. Manage all ERP data via natural conversation.

Oracle NetSuite MCP Server connects your AI agent directly to your entire Enterprise Resource Planning system. Forget clicking through menus and dashboards; you just talk to NetSuite, and your agent pulls the answers for financial data, sales orders, inventory counts, or customer details instantly. You're running full ERP management purely through conversation.

Your AI client gets access to nine specific tools that let your agent act like a trained administrator inside NetSuite. It doesn't just read data—it makes changes and runs deep reports without you having to log into the UI yourself. This is about action, not just viewing.

Running Custom Reports and Complex Queries
You can run advanced, custom queries against any structured NetSuite dataset using SQL-like commands via execute_suiteql. You write standard SELECT, JOIN, and WHERE statements; the agent executes them directly on the database. This lets you build reports that would normally take an analyst hours to pull together.

Listing and Auditing Records
To get a broad view of your business assets, you use list_records, which pulls general lists for types like customers, vendors, or inventory items so you can audit the full range of data. You'll find list_customers gives you a complete roster of every customer account stored in NetSuite.

If you need to check recent sales activity, list_sales_orders pulls a list of the most recently created or updated orders.

Retrieving Specific Data Points
When you know the exact ID of a record—say, an invoice number or a vendor ID—you use get_record to pull all available fields for that specific item. If you're building something and need to know what data is even possible on a certain record type before you try writing it, run through get_record_metadata.

This tells you exactly which fields exist on any given NetSuite record structure.

Creating and Modifying Records (Making Changes)
Your agent can generate brand new business documents using create_record, provided you give it the specific record type and all the necessary field data. You'll create things like sales orders or journal entries this way. Similarly, you use update_record to change specific fields on an existing document—you have to specify exactly what needs changing so your agent doesn't mess anything up.

For cleanup, delete_record permanently removes a record from NetSuite; remember that action is irreversible.

Putting It Together in Practice
When you use the server, your AI client sends an intent—for example, 'Show me all journal entries for Q3.' The MCP Server translates that into structured tool calls. If you need to run a report, it invokes execute_suiteql. To find specific sales data, it might call list_sales_orders and then use the results with get_record.

You can manage the entire lifecycle of an item: listing vendors via list_records, getting a specific vendor's contact details using get_record, and finally creating a new payment record using create_record. The agent handles all the back-and-forth, returning actionable data or confirmation that the transaction went through.