Platform Components

These pages describe the main technical building blocks behind the Health AI Platform.

Components

Workflow Engine

@loop/workflow-engine executes declarative workflow definitions.


import { WorkflowEngine } from '@loop/workflow-engine';
 
const workflow = WorkflowEngine.load({
  id: 'demo-v1',
  name: 'Demo',
  version: '1.0.0',
  steps: [{ id: 'check', type: 'check_biomarker' }],
});

Read Workflow Engine.

Tool Registry

Luna exposes a concrete registry of named tools and tool categories.


import { lunaTools } from '@/lib/tools';
 
console.log(Object.keys(lunaTools));

Read Tool Registry.

Research Engine

Research ingestion extracts text from PDFs, chunks content by section, and generates embeddings.


import { createResearchPaperIngestionPipeline } from '@loop/ai';
 
const pipeline = createResearchPaperIngestionPipeline({
  apiKey: process.env.OPENAI_API_KEY!,
});

Read Research Engine.

ML Layer

The Python ML service currently exposes a health route and a stable client boundary.


import { MLClient } from '@loop/ml-client';
 
const ml = new MLClient({ baseUrl: 'http://localhost:8000' });
await ml.healthCheck();

Read ML Layer.

How the pieces fit


Workflow definition
    ->
Workflow execution
    ->
Patient context and tool boundaries
    ->
Knowledge or ML enrichment
    ->
Application response