straive-prototypes

AI Prototypes

This repository contains multiple AI-driven prototype applications built for compliance engines. Reusable utilities, styles, and configurations are shared inside the /shared folder.

🏗️ Directory Structure

/prototypes
  ├── /shared/                  # Shared resources across prototypes
  │     └── /js/
  │           └── config.js     # Shared LLM endpoint and key management modal logic
  │
  ├── /agentic-nc-writer/       # NC/CAPA Writer Prototype (Aura Quality Engine)
  │     ├── index.html          # Workspace markup
  │     ├── index.css           # Workspace styling & print rules
  │     ├── app.js              # State orchestration & UI controllers
  │     ├── agents.js           # Multi-agent prompt directives
  │     ├── templates.js        # Incident scenario templates
  │     └── README.md           # Prototype-specific documentation & workflow
  │
  ├── /air-ocean-cost-converter/ # Air 2 Ocean Cost Converter Prototype
  │     ├── index.html          # Workspace markup
  │     ├── index.css           # Workspace styling & print rules
  │     ├── app.js              # Calculation engine & UI controllers
  │     └── README.md           # Prototype-specific documentation & workflow
  │
  └── README.md                 # This file (Index of all prototypes)

🚀 Active Prototypes Index

  1. Agentic NC/CAPA Writer (Aura): An automated multi-agent report generator and compliance auditing tool designed for quality excursions.
  2. Air 2 Ocean Cost Converter: An interactive cargo decision cockpit that evaluates cost-to-time freight trade-offs (including tied-up inventory carrying costs) and drafts switch proposals using a Logistics Optimizer Agent.

🏗️ System Architectures

1. Aura Quality Engine (Agentic NC/CAPA Writer)

The following block diagram outlines the data flow between the cooperative compliance agents:

graph TD
    A[Operator Raw Observations] -->|Submit| B(Orchestrator)
    B --> C[Intake Parser Agent]
    C -->|Title, Record ID, Summary| D[5-Whys RCA Agent]
    D -->|Root Cause| E[CAPA Planner Agent]
    E -->|Actions Plan| F[Compliance Auditor Agent]

    F -->|Audit Score & Logs| G{Score >= 95%?}
    G -->|No - Critique & Iterate| D
    G -->|Yes| H[Email Pitch Agent]
    H -->|Complete Draft| I[Interactive UI / PDF Print]

2. Air 2 Ocean Cost Converter

The following block diagram outlines the time-to-cost trade-off engine and Logistics Optimizer Agent flow:

graph TD
    A[User Inputs: Cargo & Transit Profile] -->|Calculate| B(Calculation Engine)
    B --> C[Air Chargeable Weight]
    B --> D[Ocean Chargeable Measure]
    C & D --> E[Direct Freight Costs]
    B --> F[Daily Inventory Carrying Cost]
    E & F --> G[True Economic Savings / Variance]
    E & F --> H[Transit Delay Breakeven Buffer]
    G --> I[SVG Stacked Bar Chart]
    H --> J[Dynamic Slider Gauge]
    G & H --> K[Logistics Optimizer Agent]
    K -->|Compile Draft| L[Confidential PDF Report]

⚙️ Running Locally

Start a local server at the root directory to preview all prototypes:

# Start server from workspace root
uv run python -m http.server 8080

Once running, navigate to: