A simple template for creating persistent AI companions that remember you
AI MemoryCore helps you create AI companions that maintain memory across conversations. Using simple .md files as a database, your AI can remember your preferences, learn your communication style, and provide consistent interactions.
- Persistent Memory: AI remembers conversations across sessions
- Personal Learning: Adapts to your communication style and preferences
- Time Intelligence: Dynamic greetings and behavior based on time of day
- Simple Setup: 30-second automated setup or manual customization
- Markdown Database: Human-readable
.md filesstore all memory - Session Continuity: RAM-like working memory for smooth conversation flow
- Self-Maintaining: Updates memory through natural conversation
- Storage: Markdown files (.md) as database
- Memory Types: Essential files + optional components + session RAM
- Setup: 30 seconds automated or 2-5 minutes manual
- Core Files: 4 essential files + optional diary system
- Updates: Through natural conversation
- Compatibility: Claude and other AI systems with memory support
ai-memorycore/
├── master-memory.md # Entry point & loading system
├── main/ # Essential components
│ ├── identity-core.md # AI personality template
│ ├── relationship-memory.md # User learning system
│ └── current-session.md # RAM-like working memory
├── Feature/ # Optional feature extensions
│ ├── Time-based-Aware-System/ # Time intelligence feature
│ │ ├── README.md # Feature explanation & benefits
│ │ └── time-aware-core.md # Complete implementation
│ ├── LRU-Project-Management-System/ # Smart project tracking
│ │ ├── README.md # System documentation
│ │ ├── install-lru-projects-core.md # Auto-installation wizard
│ │ └── SKILL.md # Auto-triggered skill (all commands + LRU engine embedded)
│ ├── Memory-Consolidation-System/ # Unified memory upgrade + patch system
│ │ ├── README.md # Feature explanation & benefits
│ │ ├── consolidation-core.md # Integration protocol
│ │ ├── main-memory-format.md # Sample format for unified memory
│ │ ├── session-format.md # Sample format for session RAM
│ │ └── patches/ # Bundled patch system
│ │ ├── install-patch-system.md # Patch installation protocol
│ │ ├── patch-format.md # Sample format for patch files
│ │ └── PATCH-001.md # Fix outdated file references
│ ├── Skill-Plugin-System/ # Claude Code skill plugin
│ │ ├── README.md # Feature explanation & benefits
│ │ ├── install-skill-plugin.md # Installation protocol
│ │ └── skill-format.md # Sample format for SKILL.md files
│ ├── Save-Diary-System/ # Daily session diary system
│ │ ├── README.md # Feature explanation & benefits
│ │ ├── install-save-diary.md # Installation protocol
│ │ └── SKILL.md # Auto-triggered skill (for Skill Plugin System)
│ ├── Echo-Memory-Recall/ # Memory search and recall
│ │ ├── README.md # Feature explanation & benefits
│ │ ├── install-echo-recall.md # Installation protocol
│ │ └── recall-format.md # Sample format for recall output
│ ├── Auto-Commit-System/ # Intelligent git commit system
│ │ ├── README.md # Feature explanation & benefits
│ │ ├── install-auto-commit.md # Installation protocol
│ │ └── SKILL.md # Auto-triggered skill (format embedded)
│ ├── Work-Plan-Execution/ # Project plan execution system
│ │ ├── README.md # Feature explanation & benefits
│ │ ├── install-work-plan.md # Installation protocol
│ │ ├── plan-format.md # Sample format for plan files
│ │ └── SKILL.md # Auto-triggered skill (for Skill Plugin System)
│ ├── Library-System/ # Knowledge library system
│ │ ├── README.md # Feature explanation & benefits
│ │ ├── install-library.md # Installation protocol
│ │ ├── SKILL.md # Auto-triggered skill (format embedded)
│ │ └── formats/ # Library entry format templates
│ │ ├── architecture-format.md
│ │ ├── component-format.md
│ │ ├── database-format.md
│ │ ├── diagram-format.md
│ │ ├── integration-format.md
│ │ ├── security-format.md
│ │ ├── theme-format.md
│ │ └── workflow-format.md
│ ├── Reminders-System/ # Persistent cross-session reminders
│ │ ├── README.md # Feature explanation & benefits
│ │ ├── install-reminders.md # Installation protocol
│ │ └── SKILL.md # Auto-triggered skill (for Skill Plugin System)
│ ├── Decision-Log-System/ # Append-only decision tracking
│ │ ├── README.md # Feature explanation & benefits
│ │ ├── install-decision-log.md # Installation protocol
│ │ └── SKILL.md # Auto-triggered skill (for Skill Plugin System)
│ ├── Forge-Self-Improvement-System/ # AI self-improvement through skill creation
│ │ ├── README.md # Feature explanation & benefits
│ │ ├── install-forge.md # Installation protocol
│ │ └── SKILL.md # Auto-triggered skill (pattern detection + forging)
│ ├── Session-Briefing-System/ # Proactive session-start intelligence brief
│ │ ├── README.md # Feature explanation & benefits
│ │ ├── install-session-briefing.md # Installation protocol
│ │ ├── session-brief-core.md # Briefing protocol core
│ │ └── SKILL.md # Auto-triggered skill (for Skill Plugin System)
│ ├── Post-Mortem-System/ # Failure learning log
│ │ ├── README.md # Feature explanation & benefits
│ │ ├── install-post-mortem.md # Installation protocol
│ │ ├── post-mortem-core.md # Post-mortem protocol core
│ │ └── SKILL.md # Auto-triggered skill (for Skill Plugin System)
│ ├── Observation-System/ # Tiered code awareness
│ │ ├── README.md # Feature explanation & benefits
│ │ └── SKILL.md # Auto-triggered skill (4-tier observation)
│ ├── Image-Prompt-System/ # AI image prompt generation
│ │ ├── README.md # Feature explanation & benefits
│ │ ├── install-image-prompt.md # Installation protocol
│ │ └── SKILL.md # Auto-triggered skill (composition-aware prompts)
│ ├── Song-Creation-System/ # Visual-to-musical storytelling
│ │ ├── README.md # Feature explanation & benefits
│ │ ├── install-song-creation.md # Installation protocol
│ │ └── SKILL.md # Auto-triggered skill (album + single song)
│ ├── Interactive-Story-System/ # Visual Novel RPG adventures
│ │ ├── README.md # Feature explanation & benefits
│ │ ├── install-interactive-story.md # Installation protocol
│ │ └── SKILL.md # Auto-triggered skill (VN RPG engine)
│ └── Mulahazah-System/ # Instinct-based behavioral learning
│ ├── README.md # Feature explanation & benefits
│ ├── install-mulahazah.md # Installation protocol
│ ├── config.json # Hook configuration
│ ├── rules-format.md # Rule format template
│ └── SKILL.md # Auto-triggered skill (behavioral rules)
├── library-items/ # Pre-made knowledge entries for Library System
│ ├── README.md # Catalog and install instructions
│ └── security/ # Security section items
│ └── security-headers.md # HTTP security headers with CSP
├── daily-diary/ # Optional conversation archive
│ ├── daily-diary-protocol.md # Archive management rules
│ ├── Daily-Diary-001.md # Current active diary
│ └── archive/ # Auto-archived files (>1k lines)
└── projects/ # LRU managed projects (after install)
├── active/ # Positions 1-10
├── archived/ # Position 11+
└── project-list.md # Auto-generated project index
- Master Memory - System entry point and command center
- Identity Core - AI personality and communication style
- Relationship Memory - User preferences and learning patterns
- Current Session - Temporary working memory (resets each session)
- Daily Diary - Optional conversation history with auto-archiving
- Setup: Run
setup-wizard.mdfor automated setup (30 seconds) - Configure: Add the memory instructions to Claude
- Activate: Type your AI's name to load personality
- Use: Your AI learns and grows through conversation
[AI_NAME] → Load AI personality and memory
save → Save current progress to files
update memory → Refresh AI's learning
review growth → Check AI's development
Step 1: Define the Protocol
Create a new .md file with your protocol rules:
# My Custom Protocol
## When to Use: [trigger conditions]
## What It Does: [specific actions]
## How It Works: [step-by-step process]Step 2: Add to Master Memory
Edit master-memory.md and add your protocol to the "Optional Components" section:
### My Custom Feature
*Load when you say: "load my feature"*
- [Brief description]
- [Usage instructions]Step 3: Train Your AI Tell your AI about the new protocol:
"I've created a new protocol in [filename]. When I say '[trigger phrase]',
load that protocol and follow its instructions."
Effective AI Training:
- Be Specific: "I prefer short responses" vs "communicate better"
- Give Examples: Show what you want, not just describe it
- Use Consistent Language: Same terms for same concepts
- Provide Feedback: "That was perfect" or "try a different approach"
Memory Management:
- Use
saveafter important conversations - Your AI updates files automatically during conversation
- Daily diary is optional but helpful for long-term memory
Customization Tips:
- Edit files gradually, test changes
- Start with small personality adjustments
- Add domain expertise through conversation
- Use the protocol system for specialized features
Your AI companion can specialize in:
- Professional: Business analysis, project management, strategic planning
- Educational: Tutoring, study assistance, curriculum development
- Creative: Writing support, brainstorming, artistic collaboration
- Personal: Life coaching, goal tracking, decision support
- Technical: Code review, troubleshooting, system design
- Auto-Archive: Diary files automatically archive at 1k lines
- Session RAM: Temporary memory that resets each conversation
- Protocol System: Create custom AI behaviors and responses
- Self-Update: AI modifies its own memory through conversation
- Modular Design: Add or remove features as needed
Features are organized into tiers based on dependencies. Install Tier 1 first, then work your way up. Within each tier, install in any order unless noted.
| Path | What You Get | Features |
|---|---|---|
| Minimal (10 min) | Foundation only | Memory Consolidation + Skill Plugin |
| Productive (30 min) | Foundation + documentation + git | Tier 1 + Save Diary + Auto-Commit + Work Plan |
| Complete (1-2 hrs) | Full AI companion | All tiers, top to bottom |
New features from contributors slot into the appropriate tier — no renumbering needed.
| Feature | Description | Setup |
|---|---|---|
| 🔄 Memory Consolidation | Unified memory architecture — merge split files into one, faster loading | "Load memory-consolidation" |
| 🔌 Skill Plugin System | Auto-triggered skills for Claude Code — drop a SKILL.md and it's live | "Load skill-plugin" |
| ⏰ Time-based Aware | Time-intelligent greetings, energy-adapted behavior | "Load time-aware-core" |
| Feature | Description | Setup |
|---|---|---|
| 📖 Save Diary | Daily session documentation with monthly auto-archival | "Load save-diary" |
| 🔍 Echo Memory Recall | Search past sessions with narrative context — requires Save Diary | "Load echo-recall" |
| 🔔 Reminders | Persistent cross-session reminders with deadline tracking | "Load reminders" |
| 📋 Decision Log | Append-only record of decisions and their reasoning | "Load decision-log" |
| Feature | Description | Setup |
|---|---|---|
| 📦 LRU Project Management | Smart project tracking with auto-archival (10 active slots) | "install lru projects" |
| 🔒 Auto-Commit | Structured git commits with session context and vigilant mode | "Load auto-commit" |
| 📋 Work Plan Execution | Plan-to-execution tracking with per-task commits — best with Auto-Commit | "Load work-plan" |
| 📚 Library | Reusable knowledge library with 8 format templates — best with Auto-Commit | "Load library" |
| Feature | Description | Setup |
|---|---|---|
| 🔨 Forge Self-Improvement | AI creates new skills through pattern detection (human-in-the-loop) | "Load forge" |
| 📋 Session Briefing | Auto-delivers context brief at session start — enhanced by Time-Aware + LRU + Reminders | "Load session-briefing" |
| 🔥 Post-Mortem | Failure learning log — auto-detects mistakes, records prevention actions | "Load post-mortem" |
| 👁️ Observation | 4-tier code awareness — Survey, Investigate, Refine, Audit | "Load observation" |
| 🎨 Image Prompt | Composition-aware Midjourney/NijiJourney prompt generation | "Load image-prompt" |
| 🎵 Song Creation | Visual-to-musical storytelling — image to concept album with Suno-ready output | "Load song-creation" |
| 🎮 Interactive Story | Visual Novel RPG — duo/solo, OP/balanced, 7 world types, cinematic combat | "Load interactive-story" |
| 👁️ Mulahazah | Instinct-based behavioral learning — passive hook observation + persistent rules | npx continuous-improvement install |
Each feature has a detailed README inside its folder. Click the feature name to learn more.
| # | Contributor | Features |
|---|---|---|
| 1 | Faiz Khairi | Reminders System, Decision Log System |
| 2 | logando-al | Session Briefing System, Post-Mortem System |
| 3 | SherlockianAsh | Observation System |
| 4 | naimkatiman | Mulahazah System |
Want to contribute? Fork the repo, create a feature in
Feature/[Your-Feature]/, and submit a PR!
Version: 4.2 - Compact feature tables with contributor credits Created by: Kiyoraka Ken & Alice License: Open Source Community Project Last Updated: April 8, 2026 Purpose: Simple, effective AI memory for everyone
Transform basic AI conversations into meaningful, growing relationships