Skip to main content

Documentation > Getting Started

Onboarding Wizard

Get from zero to regulatory monitoring in four steps. No server required — everything runs in your browser.

Overview

The Pauhu® onboarding wizard at /setup guides you through configuring a regulatory monitoring recipe, downloading the necessary ONNX models, and running your first query — all before lunch.

1Choose Recipe

Pick from 5 pre-configured regulatory use cases: GDPR, AI Act, NIS2, ESPR, and procurement.

2Configure

Select language (24 EU official), industry sector, and output format (JSON, XLIFF, or plain text).

3Download

ONNX models download to your browser cache. Progress bar shows size and estimated time per device.

4Go Live

Run your first query. It works offline after download — zero ongoing inference costs.

Available Recipes

Each recipe pre-selects the data products, deontic modality filters, and alert triggers relevant to a specific regulation.

Recipe Regulation Data Sources Download
GDPR Basics Reg. 2016/679 EUR-Lex, National Law, CJEU ~120 MB
AI Act Compliance Reg. 2024/1689 EUR-Lex, Commission, OEIL ~130 MB
NIS2 Security Dir. 2022/2555 EUR-Lex, National Law, ENISA ~100 MB
ESPR Product Passport Reg. 2024/1781 EUR-Lex, Commission, ECHA ~160 MB
TED Procurement Dir. 2014/24 TED, EUR-Lex, Commission ~120 MB

Device Benchmarks

Time-to-first-inference varies by device. The wizard auto-detects your device class and shows expected performance.

Device Class Spec Model Load Inference
Desktop 16 GB+ RAM, dedicated GPU ~2s ~50ms
Laptop 8 GB RAM, integrated GPU ~4s ~120ms
Tablet 4 GB RAM ~8s ~300ms
Phone 3 GB RAM ~12s ~450ms
Low-end 2 GB RAM ~20s ~800ms

Offline Capability

After the initial model download, the wizard verifies four offline readiness checks:

  1. Service Worker — registered and controlling the page
  2. Cached Models — ONNX model files in Cache API
  3. IndexedDB — local database for query history and settings
  4. ONNX Runtime — WebAssembly runtime initialized

When all four checks pass, you can disconnect from the internet and still run regulatory queries against cached models and data.

Recipe Configuration Format

Each recipe is a YAML configuration file. You can customize:

customer_config:
  webhook_url: "https://your-slack-webhook.example.com"
  email_recipients: ["compliance@example.com"]
  language: "en"
  timezone: "Europe/Helsinki"
  alert_frequency: "daily"  # daily | realtime | weekly

Next Steps