SANA OS: A Conversational Operating System Designed for Large Language Models

SANA OS (Structural Alignment Engine) is a conversational operating system designed specifically for large language models, with the core goal of resolving structural misalignment issues behind conflicts. It helps AI achieve true understanding in conversations rather than mere agreement through its three-layer behavioral model and multiple professional persona frameworks. The core philosophy of the project is: most conflicts stem from misaligned premises (both parties defend different survival baselines), so premises must be aligned before drawing conclusions.

Traditional AI conversational systems often pursue quick responses and superficial correctness, but ignore the structural factors behind problems. The designers of SANA OS believe that the deep root of human conflicts lies in the tension between survival needs, narrative identity, and institutional constraints. Based on this, a three-layer behavioral model was constructed:

• Layer A (Survival Layer): Basic survival needs (food, sleep, safety, etc.)
• Layer B (Narrative Layer): Identity, values, sense of belonging, etc.
• Layer C (Institutional Layer): Formal systems such as laws, policies, contracts, etc. The key insight is the Proximity Rule: Priority = Impact / Distance; people prioritize handling immediate and visible Layer A/B pressures over distant Layer C rules, which is a structural response rather than a character flaw.

The SANA OS architecture is divided into three parts:

1. Core Layer: Basic components that must be loaded, including four key files (Core_Principle.md, Core_Specification.md, Integrated_Knowledge.md, Communication_Layer.md), and defines four unmodifiable principles:
   • born_loved = true (non-hostile baseline)
   • No logic without observation (premise alignment precedes analysis)
   • Structural translation (transform "you are wrong" into structural consequences)
   • Autonomy preservation (present structures and options without forcing decisions)
2. Persona Layer: Specialized modes for different scenarios, such as Coaching (personal/workplace conflicts), History (historical event analysis), Reasoning (team decision-making), etc.
3. Framework Layer: Optional analysis tools, such as GMM (Governance Maturity Model), RBM (Resource Balance Model), CPM (Cognitive Processing Model), etc., which are selectively loaded based on the persona.

Using SANA OS requires loading files in a specific order. Taking the Coaching persona as an example:

1. core/Core_Principle.md
2. core/Integrated_Knowledge.md
3. core/Communication_Layer.md
4. frameworks/RBM.md
5. frameworks/CPM.md
6. frameworks/Value_Formation.md
7. personas/Persona_Coaching.md Users can create a project on an AI platform and paste the combined system prompt (context persists). The project provides pre-configured GPT versions and 5 printable worksheets for non-technical users.

SANA OS uses the SANA OS License v1.0 open-source agreement: commercial use is allowed, but it must be labeled "Powered by SANA OS" and linked to sana-os.org; it is forbidden for manipulation or coercion purposes; core/Core_Principle.md cannot be modified to ensure core principles are not distorted. The project has established a complete community support system, including an X (Twitter) account, contact email, and security reporting channels to facilitate continuous development.

SANA OS provides a new conversational paradigm: first understand the other party's structural situation, then identify the deep causes of differences, rather than rushing to prove correctness or simply agree/disagree. Its application scenarios include:

• Workplace mediation: Help conflicting parties see the source of pressure instead of blaming
• History education: Understand the structural factors behind historical events
• Team decision-making: Coordinate priority conflicts between departments
• Personal growth: Understand one's own behavioral patterns and value formation process In an era of information overload and polarized opinions, this method has important practical significance.

SANA OS represents an important direction in AI system design: shifting from pursuing surface intelligence to deep understanding, making AI more "wise" (identifying structural factors and promoting true communication). It is not only a technical project but also a philosophical exploration of human-AI relations—helping humans make better judgments rather than replacing them. As the capabilities of large language models improve, SANA OS is expected to play a key role in future AI applications.