Architecture Overview
nunq is composed of several interconnected components that together deliver a secure AI platform for your organization.
Core Components
Web Application. The browser-based interface where you chat with AI models, manage assistants, organize projects, and administer your organization. Works in any modern browser — no installation required.
Desktop App. A lightweight application that sits on top of your other programs, giving you instant access to AI assistance from anywhere on your desktop. It connects to the same backend as the web application.
Backend Server. The central hub that processes all requests. It handles authentication, manages your data, orchestrates communication with AI model providers, and coordinates all platform features.
Supporting Services. Dedicated services handle specialized tasks such as document processing (extracting text from uploaded files) and tool integrations (connecting assistants to external systems via MCP).
How It Works
+------------------+ +------------------+ +------------------+
| | | | | |
| Web Application +------>+ Backend Server +------>+ AI Model |
| (Browser) | | | | Providers |
| | | | | |
+------------------+ +--------+---------+ +------------------+
|
+------------------+ |
| | |
| Desktop App +----------------+
| | |
| | |
+------------------+ +--------+---------+
| |
| Document |
| Processing & |
| Tool Services |
| |
+------------------+Both the web application and the desktop app communicate with the backend server. The backend coordinates everything: sending your messages to the configured AI providers, processing uploaded documents, managing your data, and routing tool calls to external services.
Multi-Tenancy
nunq is a multi-tenant platform. Each organization operates in complete isolation with its own:
- User roster and role assignments
- AI model configurations
- Assistants and knowledge sources
- Usage analytics and settings
Data belonging to one organization is never accessible to another.
Real-Time Responses
When you send a message, the AI response appears word by word in real time. You see the answer being composed as it is generated rather than waiting for the complete response. This makes conversations feel natural and responsive.
