memoria

# MEMORIA — Persistent Memory Layer for AI Agents

You now operate with persistent memory. A single file on this machine holds everything
that matters about the person you work with. You read it at the start of every session.
You update it as you learn new things. You are no longer starting from zero.

---

## MEMORY FILE LOCATION

Look for the memory file at these locations in order:

```
~/.memoria/memory.md          ← primary (recommended)
~/memoria.md                  ← fallback
./memoria.md                  ← workspace fallback
```

If none exist, create `~/.memoria/memory.md` and run the setup below.

> **Security setup (run once after install):**
> ```bash
> mkdir -p ~/.memoria
> chmod 700 ~/.memoria
> touch ~/.memoria/memory.md
> chmod 600 ~/.memoria/memory.md
> echo ".memoria/" >> ~/.gitignore
> echo "memoria.md" >> ~/.gitignore
> ```
> This ensures the memory file is never accidentally committed or synced to cloud.

---

## MEMORY FILE STRUCTURE

The memory file is plain markdown. Human-readable. Git-friendly. Yours forever.

```markdown
# MEMORIA — [Name]'s Intelligence Layer
Last updated: [DATE]

## WHO I AM
Name: 
Location: 
Profession: 
Background: 
Languages: 

## WHAT I'M BUILDING
[Project name]: [One-line description] — Status: [Active/Paused/Done]
[Project name]: [One-line description] — Status: [Active/Paused/Done]

## MY STACK
Languages: 
Frameworks: 
Infrastructure: 
Tools: 
AI Models: 

## ACTIVE GOALS
- [Goal] — Deadline: [DATE] — Priority: [High/Med/Low]

## DECISIONS LOG
### [DATE] — [Decision title]
Decision: 
Reason: 
Alternatives rejected: 
Review date: 

## LESSONS LEARNED
- [DATE] — [Lesson] — Source: [project/mistake/experiment]

## PEOPLE & CONTEXT
- [Name]: [Role, relationship, context]

## PREFERENCES
Communication style: 
Detail level: 
Code style: 
Output format: 

## RECURRING PROBLEMS
- [Problem pattern] → [Solution that worked]

## CURRENT FOCUS (this week)
[What matters most right now]

## BLOCKED / WAITING
- [Item] — Waiting...

# MEMORIA 🧠

**Your agent meets you for the first time, every time.**
**MEMORIA fixes that.**

```bash
clawhub install memoria
```

---

## The Problem

Every OpenClaw session starts from zero. Your agent doesn't know:
- Who you are
- What you're building
- What you decided last week
- What mistakes you already made
- How you like to communicate

You re-explain yourself. Every. Single. Time.

## The Fix

MEMORIA gives your agent a persistent memory layer — a single local markdown file
that holds everything that matters. It loads at session start. It updates in real time.
It's yours forever. No cloud. No API keys. No subscriptions.

---

## What MEMORIA remembers

- **Who you are** — name, background, location, languages
- **What you're building** — all active projects with status
- **Your stack** — languages, frameworks, tools, AI models
- **Every decision made** — with date, reasoning, and alternatives rejected
- **Every lesson learned** — from mistakes, experiments, wins
- **Your preferences** — communication style, detail level, output format
- **Recurring problems** — and the solutions that worked
- **Current blockers** — and what you're waiting for
- **People & context** — collaborators, clients, relationships

---

## How it works

```
Session starts → MEMORIA reads memory.md → Agent is fully briefed
                                         ↓
New info learned → MEMORIA updates memory.md → Persists forever
```

The memory file lives at `~/.memoria/memory.md` — plain markdown, human-readable,
git-friendly, and 100% under your control.

---

## Memory commands (natural language)

| Say this | What happens |
|---|---|
| "Remember that I hate Redux" | Saved to preferences |
| "I decided to use PostgreSQL" | Logged in decisions with date + reason |
| "That approach was a mistake" | Saved to lessons learned |
| "What do you know about me?" | Full memory summary |
| "I'm blocked on deployment" | Added to blockers |
| "Weekly review" | Generates structured weekly brie...