AI

The official AI module for BoxLang

Welcome to the BoxLang AI Module. This module is a BoxLang module that provides AI capabilities to your BoxLang applications.

Providers

The following AI providers are supported:

• OpenAI

• DeepSeek

More coming soon.

Settings

Here are the settings you can place in your boxlang.json file:

{
	"modules" : {
		"bxai" : {
			// The provider of the AI: openai, google, aws, azure, deepseek
			provider = "openai",
			// The API Key for the provider
			apiKey = "",
			// The provider model to use, if any
			model = "gpt-4o-mini",
			// The provider properties according to provider, if any
			properties = {
			}
		}
	}
}

Usage

This module exposes the following BoxLang functions:

• aiChat( messages, model, struct data={}, boolean verbose=false ) : This function will allow you to chat with the AI provider and get responses back.

• aiChatAsync( messages, model, struct data={}, boolean verbose=false ) : This function will allow you to chat with the AI provider asynchronously and give you back a BoxLang Completable Future.

Arguments

• messages : The messages to chat with the AI. This is provider dependent. Please see each section for more information.

• model : The model to use for the AI provider. This is provider dependent. Please see each section for more information.

• data : The data to pass to the AI provider. This is provider dependent. Please see each section for more information.

• verbose : A flag to output verbose information about the AI chat or just the response message.

// Chat with the AI
aiChat( "What is the meaning of life?" );

---

OpenAI

The OpenAI provider will allow you to interact with the following APIs:

• Chat API - https://platform.openai.com/docs/api-reference/chat

• Image API

• Embedding API

Please see OpenAI API for more information.

aiChat() / aiChatAsync()

You can use the aiChat() function to chat with the OpenAI API. Here is more docs on this: https://platform.openai.com/docs/guides/text-generation

Messages

This can be any of the following

• A string : A message with a default role of user will be used

• A struct : A single message that must have a role and a content key

• An array of structs : An array of messages that must have a role and a content keys

// Chat with the AI
aiChat( "What is the meaning of life?" );

// Chat with the AI
aiChat( { role="developer", content="What is the meaning of life?" } );

// Chat with the AI
aiChat( [
	{ role="developer", content="Be a helpful assistant" },
	{ role="user", content="What is the meaning of life?" }
] );

Model

The supported models for OpenAI are:

• gpt-4o : The large model

• gpt-4o-mini : The more affordable but slower model

• gpt-4o-turbo : The turbo model

• Much more, look at the docs.

You can find more information here: https://platform.openai.com/docs/models

Data

This is an arbitrary structure that will be passed to the OpenAI API alongsside the top level body.

// Chat with the AI
aiChat( "What is BoxLang?", "gpt-4o-mini", { temperature=0.5, max_tokens=100 } );

Examples

Here are some examples of chatting with the AI:

aiChat( "Write a haiku about recursion in programming." );

aiChat( {
	"role": "user",
	"content": "Write a haiku about recursion in programming."
} );

aiChat( [
	{
		"role": "developer",
		"content": "You are a helpful assistant."
	},
	{
		"role": "user",
		"content": "Write a haiku about recursion in programming."
	}
] );

// Analyze an image
aiChat( [
    {
		"role": "user",
		"content": [
			{
				"type": "text",
				"text": "What is in this image?"
			},
			{
				"type": "image_url",
				"image_url": {
					"url": "https://upload.wikimedia.org/wikipedia/commons/thumb/d/dd/Gfp-wisconsin-madison-the-nature-boardwalk.jpg/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg"
				}
			}
		]
	}
] );

Tools

OpenAI tools are designed to enhance capabilities by providing access to real-time information, advanced computations, external integrations, and interactive features. This allows you to bind a function that will process the AI information for you in realtime.

tool = new bxmodules.bxai.models.Tool();

tool.setName( "get_weather" )
	.describe( "Get current temperature for a given location." )
	.describeLocation( "City and country e.g. Bogotá, Colombia" )
	.setFunc( ( location ) => {
		if( location contains "Kansas City" ) {
			return "85"
		}

		if( location contains "San Salvador" ){
			return "90"
		}

		return "unknown";
	});

result = aiChat( messages = "How hot is it in Kansas City? What about San Salvador? Answer with only the name of the warmer city, nothing else.", data = {
	tools: [ tool ]
} )

// San Salvador
println( result )