# QUAN 2.0

> QUAN 2.0 is an advanced, autonomous agentic AI operating system: the AI infrastructure that runs a company. It is a coordinated team of specialized AI agents with persistent memory, real tools and connectors, and approval-gated actions, working across web, a native Mac app, Telegram, and iMessage. It is an operator, not a chatbot. QUAN 2.0 is a proprietary, closed-source product built by InterSpace Labs (legal entity: InterSpace Distribution). Access is by request.

Give QUAN 2.0 a goal and it recalls context from durable memory, decomposes the goal into steps, delegates across specialized agents, calls real tools, verifies its own output with an adversarial critic, and returns finished work, pausing for human approval on anything consequential such as money, email, customer records, or the outside world. The model is treated as untrusted: it has no direct access to production databases or servers, code runs sandboxed behind an egress firewall, secrets live in an AES-256 vault the model can name but never read, and every action is recorded in a full audit trail.

Key facts:
- Product: QUAN 2.0, an agentic AI operating system (a Company OS).
- Vendor: InterSpace Labs (legal entity: InterSpace Distribution).
- License: Proprietary and closed-source. Access is by request.
- Channels: Web, native macOS app (signed and notarized), Telegram, iMessage.
- Positioning: an operator that does the work, not a chatbot that talks about it.

## Core capabilities
- [Capabilities](https://0000.casa/#capabilities): Persistent typed memory and learning; multi-agent teams (general, sales, CMS, research) with manager, planner, and critic roles; real tools and actions (email, calendar, invoicing, multi-engine web search, image generation, code execution, browser and computer use); MCP connectors (Notion, GitHub, any Model Context Protocol server); autonomous multi-step workflows that plan, act, verify, and self-correct; document intelligence over PDF, DOCX, XLSX, images, audio, and video; private knowledge base ingestion and retrieval; voice and transcription; offensive and defensive security tooling; runtime self-extension that writes and installs its own tools, skills, and connectors.

## The full skill stack
- [Skills](https://0000.casa/#skills): Cognition and memory; multi-agent orchestration; research and the live web; real-world actions; code and automation; build, deploy, and self-extend; MCP connectors; offensive and defensive security; and a governing operating constitution (SOUL.md, AGENT.md, USER.md, QUAN.md, SECURITY.md, PENTEST.md) hot-reloaded and enforced on every channel.

## How it works
- [How it works](https://0000.casa/#how): Recall, durable memory you own and can view, edit, or delete. Execute, plan and call the right tools and connectors, then verify before reporting. Confirm, consequential actions are staged for one-tap human approval.

## Infrastructure
- [Infrastructure](https://0000.casa/#infra): Multi-model engine routing across OpenAI, Anthropic, Gemini, and local models; a plan-then-verify reasoning core; parallel multi-agent orchestration with an adversarial critic and a strict fan-out budget; a hot-reloaded, versioned operating constitution; a sandboxed runtime behind an egress firewall; an AES-256 credential vault; and native MCP support so a new integration is configuration, not a release.

## Model connection routes
- [Models](https://0000.casa/#models): QUAN is model-agnostic and routes across many providers: a Claude Pro or Max subscription, the Claude (Anthropic) API, OpenAI, Ollama, DeepSeek, Moonshot, and Kimi, plus local and self-hosted models and any OpenAI-compatible endpoint. It also routes dedicated speech, voice, vision, and image models. These connection routes are configurable and updated on request.
- Pricing: QUAN includes an all-inclusive catalog of all major models (open-source models and frontier models like Claude Opus, Sonnet, and Haiku) under a single monthly subscription. Running your own self-hosted models or your own API keys makes QUAN fully free to use, with no subscription.

## Security and trust
- [Security](https://0000.casa/#security): Hard infrastructure boundary with no direct access to production databases or servers; human-in-the-loop approval gates on email, payments, and customer writes; sandboxed code execution behind an egress firewall; an AES-256 encrypted vault where the model sees secret names but never values; per-member scopes; and a full audit trail.

## Pages
- [Home](https://0000.casa/): Product overview, capabilities, the full skill stack, infrastructure, security, and access request.
- [About](https://0000.casa/about): Why QUAN 2.0 exists, what InterSpace Labs believes, and the principle of capability with restraint.
- [Docs](https://0000.casa/docs): The full documentation: how it works, models and routing, capabilities, multi-agent orchestration, memory, security, usage limits, and team administration.

## Full content
- [llms-full.txt](https://0000.casa/llms-full.txt): The complete plain-text content of the site in one file for language models.

## Contact
- Access requests and inquiries: eric@interspacemusic.com
- Twitter / X: https://x.com/0x_dxv