---
type: concept
title: 'Production vs Demo: What''s Real'
aliases:
  - production vs demo
  - dual-mode architecture
  - what's real
  - fixture-based
modified: 2026-04-11
tags:
  - production-readiness
  - trust
  - evaluation
  - cto
  - adoption-case
---

# Production vs Demo: What's Real

> The honest inventory: SuiteCentral 2.0 has a dual-mode architecture where production AI and connectors coexist with demo/fixture modes. Here's what's real, what's demo, and why the dual-mode design is a strength, not a weakness.

## Why this page exists

The #1 question any technical evaluator asks: **"Is this real production software, or is it a demo that looks like production software?"** The answer for SuiteCentral 2.0 is: both, intentionally, and here's the transparent breakdown.

Per [[sources/tier1-repo-evaluation-batch]] (PRODUCTION-VS-DEMO-GUIDE.md, April 9 2026):

> *"The presence of demo/fixture modes is NOT a sign of incomplete implementation. It is a production-grade architectural pattern designed for enterprise flexibility."*

The dual-mode design enables: immediate exploration without API keys, cost-effective demos for sales/training, offline/air-gapped deployments, graceful degradation when services are unavailable, and A/B testing between providers.

## AI providers — all production-ready

| Provider | Production Status | Demo Available | Badge Color |
|---|---|---|---|
| **OpenAI GPT-4o / GPT-4o-mini** | ✅ Production | Yes (mock sim) | 🟢 Green |
| **Claude 4.5 Sonnet** | ✅ Production | Yes (mock sim) | 🟢 Green |
| **OpenRouter (multi-model)** | ✅ Production | N/A | 🟢 Green |
| **LMStudio (local)** | ✅ Production | N/A (always local) | 🟢 Green |
| **Rule-Based Engine** | ✅ Always available | N/A | 🔵 Blue |
| **Mock providers** | Demo only | Always available | 🟡 Yellow |

**Claude model update**: the production environment shows `claude-sonnet-4-5-20250929` — this is **Claude 4.5 Sonnet**, not the Claude 3.5 Sonnet cited in earlier sources. Model upgrade.

## Connectors — the honest inventory

| Connector | Status | Auth | Notes |
|---|---|---|---|
| **NetSuite** | ✅ Sandbox-Tested | OAuth 1.0a | Full CRUD, Squire sandbox `TSTDRV2698307` verified |
| **Salesforce** | ✅ Production | OAuth 2.0 | Full CRUD (leads, accounts, contacts, opportunities, cases) |
| **Business Central** | ✅ Production | OAuth 2.0 | OData v4 with metadata-driven field catalog |
| **HubSpot CRM** | ✅ Production | OAuth 2.0 | Contacts, Companies, Deals, Tickets (Phase 3, Jan 2026) |
| **ShipStation 3PL** | ✅ Production | API key | Order fulfillment, shipments, tracking (Phase 3, Jan 2026) |
| **Oracle ERP** | ⚠️ **Beta** | Basic Auth, ORDS | Basic CRUD only; needs production hardening |
| **20+ more** | 🏗️ Fixture-Based | Varies | Real API implementations exist; expose test data when `DEMO_MODE=1` |

**Important**: the canonical wording says "6 production connectors" but CONNECTOR_STATUS.md marks Oracle as **Beta**, not production. Honest count: **5 production + 1 beta + 10 fixture-based + 200+ planned**. The "6 production connectors demonstrated" phrasing in [[pages/concepts/canonical-metrics]] overstates Oracle's readiness. Flagged.

**"Fixture-based" ≠ "not implemented"**: every fixture-based connector has a real API implementation. The fixture mode activates only when `DEMO_MODE=1`. In production mode, the same code calls the real API.

## Features — what's production-ready

| Feature | Status | Notes |
|---|---|---|
| AI Field Mapping | ✅ Production | Real AI with OpenAI/Claude/OpenRouter/LMStudio OR rule-based fallback |
| Integration Wizard | ✅ Production | Orchestrates real connectors OR fixtures |
| AI Configuration Dashboard | ✅ Production | Manages real AI provider credentials |
| Enterprise Features (BC Metadata, Delta Sync, etc.) | ✅ Production | Real implementations |
| RAG Knowledge Base | ✅ Production | **7,415 chunks indexed** in real vector store |
| Help Chat Widget | ✅ Production | Real RAG-powered assistance |
| Telemetry & Monitoring | ✅ Production | Real metrics collection |
| Mock Dashboard APIs | 🧪 Demo Only | `src/routes/mockDashboardAPIs.ts` |

## Why this matters for the evaluation

A CTO ([[pages/entities/reuben-cook|Reuben Cook]]) evaluating the repo will see fixture files, mock providers, and demo-mode flags in the code. Without this document, those artifacts look like "it's not real." WITH this document, they're understood as a deliberate enterprise pattern: **the same codebase serves both production customers and zero-cost demo environments, controlled by environment flags, with color-coded badges so users always know which mode they're in.**

This is the document Reuben should read FIRST, before opening any source file.

## Sources

- [[sources/tier1-repo-evaluation-batch]] — PRODUCTION-VS-DEMO-GUIDE.md and CONNECTOR_STATUS.md