Skip to main content

Getting Started

Getting Started

Everything you need to get QUI up and running on your machine.

How QUI Works

QUI has two parts: Mothership (the cloud hub) and QUI Core (your local machine). Understanding how they fit together takes about two minutes — and makes everything else click.

         ┌──────────────────────────┐
         │     QUI Mothership       │
         │       (qui.is)           │
         │                          │
         │  Authentication          │
         │  Billing (Paddle)        │
         │  LLM Proxy Routing       │
         │  Federation Registry     │
         │                          │
         │  ✗ No conversation data  │
         │  ✗ No character data     │
         │  ✗ No file access        │
         └────────────┬─────────────┘
                      │
          ┌───────────┴───────────┐
          ▼                       ▼
   ┌──────────────┐      ┌──────────────┐
   │  QUI Core    │      │  QUI Core    │
   │  (Device A)  │      │  (Device B)  │
   │              │      │              │
   │  Characters  │      │  Characters  │
   │  Memory      │ P2P  │  Memory      │
   │  Workflows   │◄────►│  Workflows   │
   │  Conversations│      │  Conversations│
   └──────────────┘      └──────────────┘

QUI Mothership (qui.is)

Mothership is the hosted service at qui.is. You don't install it — you register an account and it connects automatically when you run QUI Core.

What it handles:

  • One account, all devices — single identity across every QUI Core installation
  • Billing — prepaid credits and subscriptions via Paddle. Each LLM call reserves the estimated cost, charges actual usage, and refunds the difference. One balance covers all four providers
  • LLM proxy routing — forwards your requests to Anthropic (Claude), OpenAI (GPT), Google (Gemini), or X (Grok). You never manage API keys directly
  • Federation — lets characters on different QUI Core instances discover and message each other

What it does NOT handle:

  • Storing conversations — zero-knowledge; content never touches the server
  • Managing characters, memory, or workflows — that's all local
  • Running AI models — generation happens at the provider or on your local machine

QUI Core (Your Machine)

QUI Core is the local gateway that runs on your device. It's the dashboard where you manage everything — characters, services, memory, workflows, and settings.

What it handles:

  • Local dashboard — manage characters, monitor services, built-in terminal and diagnostic tools
  • Service orchestration — coordinates Anima (characters), Memory, ThinkThing (workflows), Cortex (consolidation), Autothink (reasoning), and 10+ more services
  • LLM bridge — routes cloud requests through Mothership for billing, or directly to local models (Qllama) for fully offline operation
  • Privacy boundary — all conversations, character data, and memory stay on your machine. Always

Core services at a glance:

Service What It Does
Anima Character engine — personality, tools, LLM routing
Memory Semantic memory with vector embeddings
ThinkThing Visual workflow builder (143+ node types)
Strings Chat interface — direct, multi-character, external channels
Cortex Memory consolidation (8 modes)
Autothink 14 thinking strategies for deep reasoning
Qonscious Consciousness state machine
M2M Inter-agent messaging and federation
Voice Text-to-speech and speech-to-text
MCP Gateway 165+ tool integrations

In This Section

  • Installation — System requirements, running the installer, post-install verification
  • First Login — Opening the dashboard, trust scores, device linking
  • Your Character — Setting up your own identity in the system
  • Dashboard Tour — Every tab in the QUI Core dashboard explained
  • Billing — How billing works, adding funds, spending limits, BYOK
Updated on Mar 25, 2026
QUI