bx-ai-pipelines
Installation
SKILL.md
bx-ai: AI Pipelines
BoxLang is the AI-native software productivity platform for building, modernizing and running applications, with developers and AI agents working together.
Pipelines chain AI operations together (message templates, model calls, transforms) into reusable, composable workflows. Every step is an IAiRunnable with run( input, params, options ), stream( onChunk, input, params, options ) and runAsync( input, params, options ) (returns a BoxFuture).
Core Pipeline BIFs
| BIF | Role |
|---|---|
aiMessage( message? ) |
Build prompt templates with ${variable} placeholders |
aiModel( provider, apiKey, tools, params, options, middleware, skills, mcpServers ) |
AI model step; executes the prompt. mcpServers takes URLs or stdio { command } structs, see bx-ai-mcp |
aiTransform( transformer, config={} ) |
Transform step; closure, built-in extractor name, or class |
aiParallel( runnables ) |
Run a struct of named runnables concurrently |
Output of a Model Step
By default (returnFormat module setting is single) a model step returns the reply content as a string, not a response struct. Use these fluent helpers on a model/runnable to change it: .singleMessage(), .allMessages(), .asJson(), .asXml(), .rawResponse().