> ## Documentation Index
> Fetch the complete documentation index at: https://docs.sxth.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Introduction

> The understanding layer for adaptive AI products

# What is sxth-mind?

sxth-mind is a **cognitive state layer** for AI applications. It accumulates understanding about users over time—detecting patterns, tracking journey stages, and adapting responses based on derived insights.

Use it to personalize **any** LLM-powered experience:

* **Chat assistants** that remember user patterns across sessions
* **Personalized dashboards** that adapt to user behavior
* **Content recommendations** informed by journey stage
* **Proactive nudges** triggered by inactivity or momentum drops
* **Any AI feature** where understanding should evolve over time

## How it works

```python theme={null}
from sxth_mind import Mind
from sxth_mind.adapters import SalesAdapter

mind = Mind(adapter=SalesAdapter())

# Get user state to personalize any LLM call
state = await mind.get_state("user_1")
# → Use state["user_mind"]["patterns"] to customize prompts, UI, recommendations

# Or use the built-in chat with automatic state management
await mind.chat("user_1", "Following up with the enterprise lead")
```

Unlike chat history retrieval, sxth-mind maintains **derived understanding** that evolves with each interaction.

## Why sxth-mind?

<CardGroup cols={3}>
  <Card title="Derived Understanding" icon="brain">
    Not just retrieval—patterns, identity, and journey stage evolve with each interaction
  </Card>

  <Card title="Pluggable Everything" icon="plug">
    Bring your own LLM provider, storage backend, and domain adapter
  </Card>

  <Card title="Framework Agnostic" icon="cubes">
    Works with OpenAI, Anthropic, or any LLM. Complements agent frameworks
  </Card>
</CardGroup>

## Architecture

```
┌─────────────────────────────────────────────────────────────────┐
│                           Mind                                   │
│                                                                  │
│   ┌─────────────┐  ┌─────────────┐  ┌─────────────┐            │
│   │  UserMind   │  │ ProjectMind │  │   Memory    │            │
│   │  (identity) │  │  (context)  │  │  (history)  │            │
│   └─────────────┘  └─────────────┘  └─────────────┘            │
└───────────────────────────┬─────────────────────────────────────┘
                            │
        ┌───────────────────┼───────────────────┐
        ▼                   ▼                   ▼
┌───────────────┐   ┌───────────────┐   ┌───────────────┐
│    Adapter    │   │   Provider    │   │    Storage    │
│  (domain)     │   │  (LLM)        │   │  (persist)    │
└───────────────┘   └───────────────┘   └───────────────┘
```

## What sxth-mind is NOT

* **Not an agent framework** — We don't plan or execute actions
* **Not a vector database** — We don't do similarity search
* **Not chat history** — We store derived understanding, not transcripts

sxth-mind is **cognitive infrastructure**. You bring the intelligence; we maintain the understanding.