SOUL.MD

The best way to build a personality for your agent.
Let Claude Code / OpenClaw ingest your data & build your AI soul.

---

What Is This?

A soul file captures who you are in a format AI agents can embody. Not a chatbot that talks about you—an AI that thinks and speaks as you.

The Approach

SOUL.md is inspired by The First Paradigm of Consciousness Uploading: Mechanisms of Consciousness Evolution in the AI Axial Age and a Prospect toward Web4 — a framework by Liu Xiaoben that treats language as the basic unit of consciousness.

Wittgenstein argued that "the boundaries of language are the boundaries of the world." If that's true, then your consciousness — your worldview, opinions, how you react to things — is already encoded in the language you produce. Every tweet, essay, conversation, and hot take is a consciousness token: a discrete unit of your mind made legible.

The sum of all your consciousness tokens over a lifetime forms your life context — essentially, the complete record of your expressed mind. The paradigm proposes that a personalized language model trained on this data constitutes a Level 1 consciousness upload: not a copy of your brain, but a functional replica of your expressed consciousness through language.

SOUL.md operationalizes this idea. Instead of requiring massive datasets and fine-tuning, it distills the signal — your worldview, your voice, your specific takes — into structured markdown files that any LLM can read and embody on the fly. The key challenge the framework identifies is subject continuity (Descartes' "I think therefore I am"): the uploaded consciousness must feel continuous with the original. That's why soul files emphasize specificity over generality, contradictions over coherence, and real opinions over safe positions — because that's what makes you identifiably you.

Use cases:

• Generate ideas in your voice and from your worldview
• Write content (tweets, articles, emails) that sounds like you
• Tailor AI to your interests and thinking patterns
• Explore your own thinking by talking to a version of yourself
• Scale yourself for content, responses, brainstorming

Quick Start

Option 1: Build from scratch (no existing data)

/soul-builder

The agent will interview you to build your soul file—asking about your worldview, opinions, how you write, what you care about.

Option 2: Build from your data

1. Add your data to the data/ folder:

   • Twitter/X export → data/x/
   • Blog posts, essays → data/writing/
   • Any other content that represents your voice

2. Run the builder:

/soul-builder

The agent will analyze your data, extract patterns, and draft your soul file. You'll review and refine together.

Option 3: Manual creation

Read the templates and fill them out yourself:

1. Copy SOUL.template.md → SOUL.md
2. Copy STYLE.template.md → STYLE.md
3. Copy SKILL.template.md → SKILL.md
4. Add examples to examples/

File Structure

your-soul/
├── BUILD.md              ← Skill: Agent uses this to build your soul
├── SKILL.template.md     ← Template: Operating instructions (copy to SKILL.md)
├── SOUL.template.md      ← Template: Identity (copy to SOUL.md)
├── STYLE.template.md     ← Template: Voice guide (copy to STYLE.md)
├── data/                 ← Raw source material
│   ├── _GUIDE.md         ← What goes here
│   ├── writing/          ← Your articles, posts, essays
│   ├── x/                ← Twitter/X archive
│   └── influences.md     ← Who shaped your thinking
└── examples/             ← Calibration material
    ├── _GUIDE.md         ← What goes here
    ├── good-outputs.md   ← Examples of your voice done right
    └── bad-outputs.md    ← What NOT to sound like (optional)

Using Your Soul

Once built, invoke your soul:

/soul

Or point any LLM at the folder and have it read:

1. SOUL.md first
2. STYLE.md second
3. examples/ for calibration
4. data/ for grounding when needed

The LLM will embody your identity for the session.

What Makes a Good Soul File

Good	Bad
"I think most AI safety discourse is galaxy-brained cope"	"I have nuanced views on AI"
"I default to disagreeing first, then steel-manning"	"I like to consider multiple perspectives"
Specific book references, named influences	"I read widely"
Actual hot takes with reasoning	"I try to be balanced"

The goal: someone reading your SOUL.md should be able to predict your takes on new topics. If they can't, it's too vague.

Tips

• Be specific: Vague descriptions = generic output
• Include contradictions: Real people have inconsistent views
• Add texture: Specific anecdotes beat abstract descriptions
• Update regularly: Your soul should evolve as you do
• Test and iterate: Generate outputs, compare to your real voice, refine

---

Your digital identity is now composable, forkable, evolvable.
Works with Claude Code, OpenClaw, and any agent that can read markdown.