llm.generateText Syntax and Usage for NetSuite AI

TL;DR: The llm.generateText method enables developers to interact with AI models effectively in SuiteScript, providing flexible parameters for customized text generation.

Supported AI Models

The llm.generateText(options) and llm.generateTextStreamed(options) methods support various models, each with specific capabilities. The table below summarizes the models and their features:

Syntax Overview

The syntax for using the llm.generateText method is straightforward, enabling developers to create AI-enhanced applications easily. Below is a sample code snippet demonstrating its basic structure:

1// Add additional code
2...
3 
4const response = llm.generateText({
5   prompt: "Hello World!",
6   modelFamily: llm.ModelFamily.COHERE_COMMAND,
7   ...
8});
9 
10...
11// Add additional code

Note: This example illustrates the syntax but does not represent a complete functional script. For more comprehensive examples, consult the N/llm Module Script Samples.

Important Parameters

The llm.generateText method accepts various parameters that customize its behavior:

• options.prompt: Required if options.toolResults is not specified. This is the prompt for the LLM.
• options.chatHistory: Optional chat history for context.
• options.documents: Provides additional documents for context when using Cohere models.
• options.modelFamily: Specifies which LLM to use, defaulting to the Cohere Command model if not set.
• options.modelParameters: Additional parameters to customize model behavior.

Who This Affects

• Developers: Those utilizing SuiteScript to interface with AI models will find this information vital for implementation.
• Administrators: Understanding the capabilities of AI models can assist in configuration and management of SuiteScript resources.

Key Takeaways

• The llm.generateText method enables interaction with various AI models in SuiteScript.
• Provides flexible parameters for tailored AI responses.
• Incorporates features like Retrieval-augmented generation (RAG) and preambles for enhanced context.
• Supports multiple model families, particularly useful for developers aiming to optimize AI-driven functionalities.