Bx ai chatting

Skill ortus-boxlang/skills/boxlang-modules/bx-ai/bx-ai-chatting

BoxLang AI skills repository and Claude Plugin

Install

npx -y skills add ortus-boxlang/skills --skill bx-ai-chatting

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Use this skill when writing BoxLang AI chat code: aiChat(), aiChatAsync(), aiChatStream(), parameters (temperature, max_tokens, model), provider selection, API keys, return formats, multi-turn conversations, and error handling.

SKILL.md

4.0 KB, as published. Nobody here has run it

bx-ai: Chatting with AI

Core BIF: `aiChat()`

// Signature
aiChat( message, params={}, options={} )

message — string or array of message structs
params — model parameters (temperature, max_tokens, model, top_p, stop, etc.)
options — provider options (provider, apiKey, returnFormat, timeout)
Returns: string by default; array or struct depending on returnFormat

Simple Usage

// Simplest call — uses configured default provider
answer = aiChat( "What is the capital of France?" )

// With parameters
code = aiChat(
    "Write a Fibonacci function in BoxLang",
    { temperature: 0.2, max_tokens: 500 }
)

// With a specific provider
result = aiChat(
    "Summarize this text: ...",
    { temperature: 0.5 },
    { provider: "claude" }
)

Parameters Reference

params = {
    temperature : 0.7,      // 0.0 (deterministic) → 1.0+ (creative)
    max_tokens  : 1000,     // max response length
    model       : "gpt-4o", // provider-specific model name
    top_p       : 1.0,      // nucleus sampling (use OR temperature, not both)
    stop        : ["\n\n"]  // stop sequences
}

Temperature guide:

0.0–0.3 — Facts, code, data extraction
0.5–0.7 — General chat, balanced
0.8–1.0 — Creative writing, brainstorming

Return Formats

// Default: returns string
text = aiChat( "Hello" )

// Full response struct (includes usage, model, finish_reason, etc.)
response = aiChat( "Hello", {}, { returnFormat: "full" } )
println( response.content )      // the AI text
println( response.usage.total )  // tokens used

// Multiple choices/candidates
choices = aiChat( "Tell a joke", { n: 3 }, { returnFormat: "choices" } )

Multi-Turn Conversations

// Pass an array of message structs
messages = [
    { role: "system",    content: "You are a helpful assistant." },
    { role: "user",      content: "What is 2+2?" },
    { role: "assistant", content: "4" },
    { role: "user",      content: "Multiply that by 10." }
]

result = aiChat( messages )
// "40"

Async Chat

// Non-blocking — returns a BoxFuture
future = aiChatAsync( "Explain quantum entanglement" )

// Do other work here...

response = future.get()    // blocks until complete
// or with timeout:
response = future.get( 30, "seconds" )

Streaming Responses

// aiChatStream — provides real-time chunks
aiChatStream(
    "Write a short story about a robot",
    {},
    {},
    ( chunk ) -> {
        // called for each token chunk
        print( chunk )
    }
)

Provider Configuration

// Provider in options
result = aiChat( "Hello", {}, { provider: "openai", apiKey: "sk-..." } )
result = aiChat( "Hello", {}, { provider: "claude", apiKey: "sk-ant-..." } )
result = aiChat( "Hello", {}, { provider: "gemini" } )
result = aiChat( "Hello", {}, { provider: "ollama" } )   // local, no key needed

// Available providers: openai, claude, gemini, grok, groq, deepseek, ollama, mistral

Error Handling

try {
    result = aiChat( "Hello", {}, { provider: "openai" } )
} catch ( bxModules.bxai.exceptions.AIProviderException e ) {
    // Provider-level error (bad key, rate limit, etc.)
    logError( "AI provider error: #e.message#" )
} catch ( bxModules.bxai.exceptions.AITimeoutException e ) {
    // Response took too long
    return getDefaultResponse()
}

Common Pitfalls

❌ Do NOT pass model: as a top-level BIF argument — put it in params
❌ Do NOT use both temperature and top_p simultaneously
❌ Do NOT forget to handle provider exceptions in production code
✅ Set returnFormat: "full" when you need token counts or finish reasons
✅ Use aiChatAsync() for long-running requests in web handlers