Project Memory
An intelligent personal CRM that processes WhatsApp conversations to build a searchable knowledge base about contacts using diarization, transcription, and PII sanitization. It exposes MCP tools for semantic search, contact summaries, and reminder management within Claude Desktop.
README
Project Memory 🧠
CRM personal inteligente que procesa conversaciones de WhatsApp para construir conocimiento sobre tus contactos, expuesto vía Model Context Protocol (MCP).
✨ Características
- Procesamiento de Audio: Diarización (pyannote) + Transcripción (Groq Whisper)
- Identificación de Hablantes: Distingue tu voz de la del contacto
- Privacidad: Sanitización de PII con Microsoft Presidio (RUT, teléfonos, emails)
- Extracción Inteligente: Qwen 3 32B via Groq para extraer hechos, sentimientos, temas
- Búsqueda Semántica: pgvector para encontrar conversaciones por contexto
- MCP Server: Expone tools para usar con Claude Desktop
🏗️ Arquitectura
WhatsApp → WaHA → n8n → Redis → Worker Python → PostgreSQL
↓
Claude Desktop ← MCP Server
🚀 Quick Start
# 1. Clonar
git clone https://github.com/crtormo/project-memory.git
cd project-memory
# 2. Configurar
cp .env.example .env
# Editar .env con tus API keys
# 3. Levantar
docker-compose up -d
# 4. Vincular WhatsApp
# Abrir http://localhost:3000 y escanear QR
🔧 Configuración
Variables de Entorno
| Variable | Descripción |
|---|---|
HUGGINGFACE_TOKEN |
Token para pyannote (diarización) |
GROQ_API_KEY |
API key de Groq (Whisper + Qwen) |
GOOGLE_API_KEY |
API key de Google AI Studio (fallback) |
Servicios Docker
| Servicio | Puerto | Descripción |
|---|---|---|
| PostgreSQL | 5432 | Base de datos + pgvector |
| Redis | 6379 | Cola de mensajes |
| WaHA | 3000 | API WhatsApp |
| n8n | 5678 | Automatización |
🛠️ Herramientas MCP
| Tool | Descripción |
|---|---|
get_contact_summary |
Resumen completo de un contacto |
query_conversations |
Búsqueda semántica en historial |
add_reminder |
Crear recordatorio |
get_pending_reminders |
Ver recordatorios próximos |
list_contacts |
Listar todos los contactos |
Configurar Claude Desktop
Agregar a ~/.claude/config.json:
{
"mcpServers": {
"project-memory": {
"command": "docker",
"args": ["exec", "-i", "project-memory-mcp", "python", "-m", "src.mcp_server.server"]
}
}
}
📁 Estructura
project-memory/
├── src/
│ ├── core/ # AudioProcessor, PrivacyService, Intelligence
│ ├── database/ # Models, Repositories, Connection
│ ├── services/ # BatchProcessor, MessageQueue, Notifications
│ └── mcp_server/ # Servidor MCP con tools
├── scripts/ # CLI utilities
├── docker/ # Dockerfiles
└── n8n/ # Workflows
📊 Stack Tecnológico
- Python 3.11 + Poetry
- PostgreSQL 16 + pgvector
- Redis 7 para cola de mensajes
- pyannote.audio para diarización
- Groq API (Whisper + Qwen 3 32B)
- Microsoft Presidio para PII
- MCP SDK para Model Context Protocol
📝 Scripts CLI
# Verificar BD
python scripts/init_db.py
# Registrar tu voz
python scripts/enroll_voice.py
# Exportar datos
python scripts/export_data.py contacts --format csv
# Ver recordatorios
python scripts/check_reminders.py
🔒 Seguridad
- Las API keys nunca se commitean (están en
.env) - PII sanitizado antes de persistir
- Datos almacenados localmente (Home Lab)
📄 Licencia
MIT
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.