Mémoire

A JSON-enhanced universal tool for preserving AI conversations and development insights across any project.

🎯 Problem Solved

Never lose AI conversations again! This tool provides:

• Persistent memory - All conversations saved locally per project
• Searchable history - Find past solutions quickly
• Project organization - Each codebase maintains its own investigation database
• Zero complexity - Simple SQLite database with Make commands

🚀 Quick Setup

Method 1: One-Line Install

curl -sSL https://raw.githubusercontent.com/evoludigit/memoire/main/install.sh | bash

Method 2: Manual Setup

# Download the tool
wget https://raw.githubusercontent.com/evoludigit/memoire/main/investigation_db.py

# Initialize in your project
python3 investigation_db.py setup

Method 3: Clone and Copy

# Clone the tool repository
git clone https://github.com/evoludigit/memoire.git

# Copy to your project
cp memoire/investigation_db.py your-project/
cd your-project
python3 investigation_db.py setup

Method 4: Local Development Setup

# If you have the tool in ~/code/
cp ~/code/memoire/investigation_db.py .
python3 investigation_db.py setup

💡 Daily Usage

# Start investigating an issue
make investigation-start
# Enter: "Debug webhook deployment issue"

# During work, capture key insights
python3 -c "from investigation_db import capture_conversation; capture_conversation('Why webhook failing?', 'Found GitHub Actions only triggers on main branch')"

# Search past solutions
make investigation-search
# Enter: "webhook"

# Complete when solved
make investigation-complete
# Enter: "Fixed by creating dev-specific workflow"

🔍 Features

• JSON-First Storage - AI-friendly structured data format
• Auto-detection - Automatically detects project name from git
• Local storage - SQLite database per project with JSON documents
• Make integration - Ready-to-use Make commands
• Full-text search - Advanced search across all conversations
• Rich metadata - Captures context, statistics, and project information
• Easy export - Complete JSON export for sharing/backup
• AI integration - Includes CLAUDE.md instructions

📁 Files Created

• investigation_db.py - Main tool (700+ lines, JSON-enhanced)
• memoire_projectname.db - SQLite database with JSON documents
• Makefile commands - Investigation commands
• .claude/CLAUDE.md - Instructions for AI assistants

🆕 JSON Enhancements

AI-Optimized Data Format

{
  "investigation": {
    "uuid": "abc-123",
    "title": "Debug webhook issue",
    "status": "active",
    "project_context": {
      "git": {"current_branch": "dev"},
      "timestamp": "2025-09-23T11:42:04Z"
    },
    "statistics": {
      "message_count": 5,
      "action_count": 3,
      "finding_count": 2
    }
  },
  "conversation_entries": [
    {
      "role": "human",
      "content": "Why isn't the webhook triggering?",
      "message_type": "question",
      "char_count": 33,
      "word_count": 6,
      "metadata": {"context": "debugging"}
    }
  ]
}

New JSON Commands

python3 investigation_db.py current-json    # Full investigation as JSON
python3 investigation_db.py export-json     # Export all data as JSON

🎨 Integration Examples

Capture Conversations in Code

from investigation_db import capture_conversation

def debug_api():
    # ... debugging logic ...
    capture_conversation(
        "API returning 500 errors",
        "Found issue in database connection pool - fixed by increasing max_connections"
    )

Enhanced Commands

# Run tests with automatic logging
make investigation-test

# Git commands with logging
make investigation-git-status

🏆 Benefits

• ✅ Never lose solutions - Everything preserved locally
• ✅ Quick problem solving - Search past work instantly
• ✅ Team knowledge - Share databases across team
• ✅ Learning acceleration - Build on previous discoveries
• ✅ Zero overhead - Simple tools, no complex setup

Perfect for any development project using AI assistants!