Espanol

Que es Auto-CRM?

Auto-CRM es un CRM open-source que se personaliza automaticamente a tu negocio. Abrelo con Claude Code, ejecuta /setup, y el CRM se adapta a tu industria, tu pipeline, y tu forma de trabajar. Todo corre en tu computadora — tus datos nunca salen de tu maquina.

Por que Auto-CRM?

• 100% local — Base de datos SQLite en tu computadora. Cero servicios externos. Tus datos son tuyos.
• Se personaliza solo — Ejecuta /setup y Claude adapta el pipeline, fuentes de leads, idioma y tema a tu negocio.
• IA incluida — Clasifica leads, analiza tu pipeline, sugiere proximos pasos. Sin necesidad de API key.
• Gratis para siempre — Open source. Sin suscripciones. Sin limites de contactos o deals.
• Listo en 3 comandos — Clona, instala, ejecuta. Tu CRM funcionando en menos de 2 minutos.

Inicio Rapido

git clone https://github.com/Hainrixz/auto-crm.git
cd auto-crm && npm install
npm run dev

Abre http://localhost:3000 — tu CRM esta listo.

# Opcional: cargar datos demo para explorar
npm run init:seed

Personalizar con Claude Code

Abre el proyecto con Claude Code y escribe:

/setup

El asistente te pregunta:

1. Tipo de negocio e industria
2. Etapas de tu pipeline de ventas
3. De donde vienen tus leads
4. Idioma y tema visual

Y personaliza todo automaticamente.

Funcionalidades

Dashboard

Panel principal con KPIs en tiempo real: contactos totales, deals activos, valor en pipeline, leads calientes. Graficos de pipeline y actividad reciente.

Pipeline Kanban

Tablero visual drag & drop. Arrastra deals entre etapas. Cada tarjeta muestra valor, contacto y temperatura del lead.

Gestion de Contactos

Tabla con busqueda, filtros por temperatura (frio/tibio/caliente) y fuente. Score de cada lead. Click para ver detalle completo con historial.

Acciones Rapidas

Botones de WhatsApp, llamada, y copiar directamente en cada contacto. Un click para abrir chat de WhatsApp o iniciar llamada.

Seguimiento de Actividades

Timeline de todas las interacciones: llamadas, emails, reuniones, notas. Sistema de follow-ups con alertas para seguimientos vencidos.

Clasificacion de Leads

Dos modos: reglas automaticas (sin API key) o IA con Claude (opcional). Score de 0-100 y temperatura automatica.

Webhook

Recibe leads automaticamente desde formularios web (Typeform, Tally, Google Forms, Zapier). Soporta campos en espanol e ingles.

Exportacion CSV

Descarga tus contactos y deals como CSV con un click. Compatible con Excel.

Notificaciones

Banner en el dashboard para seguimientos vencidos. Notificaciones del navegador opcionales cada 5 minutos.

Email Digest

Resumen diario por correo con seguimientos pendientes, leads calientes y metricas. Requiere Resend (gratis).

Comandos de Claude Code

Integraciones

Webhook — Recibir leads automaticamente

curl -X POST http://localhost:3000/api/webhook \
  -H "Content-Type: application/json" \
  -d '{"name":"Juan","email":"[email protected]","phone":"555-1234","company":"Mi Empresa"}'

Soporta campos en espanol (nombre, correo, telefono, empresa) y formato Typeform anidado.

Exportar datos

# Descargar contactos como CSV
curl http://localhost:3000/api/export?type=contacts -o contactos.csv

# Descargar deals como CSV
curl http://localhost:3000/api/export?type=deals -o deals.csv

MCP — Claude Desktop / Claude.ai

Conecta tu CRM directamente a Claude Desktop:

{
  "mcpServers": {
    "auto-crm": {
      "command": "npx",
      "args": ["tsx", "/ruta/a/auto-crm/mcp/crm-server.ts"]
    }
  }
}

Ahora puedes decirle a Claude: "Muestrame mis leads calientes" o "Agrega un contacto nuevo" desde cualquier chat.

Stack Tecnico

Despliegue

Opcion 1 — Desarrollo local

npm run dev

Opcion 2 — Produccion local

npm run local

Opcion 3 — Docker

docker compose up -d

Opcion 4 — MCP (Claude Desktop)

Agrega la configuracion MCP mostrada arriba a tu claude_desktop_config.json.

Estructura del Proyecto

auto-crm/
├── CLAUDE.md                # Instrucciones para Claude Code
├── .claude/commands/        # 8 comandos interactivos
├── mcp/crm-server.ts        # Servidor MCP (10 herramientas)
├── scripts/init.ts          # Inicializacion de base de datos
├── src/
│   ├── app/                 # Paginas y 13 API routes
│   ├── components/          # 28 componentes React + 21 shadcn/ui
│   ├── db/                  # Schema SQLite + Drizzle ORM
│   ├── lib/                 # Utilidades (scoring, AI, constants)
│   └── types/               # TypeScript types
├── data/crm.db              # Base de datos (auto-generada)
├── Dockerfile               # Contenedor Docker
└── docker-compose.yml       # Docker Compose

Variables de Entorno

Todas son opcionales. El CRM funciona completamente sin ninguna.

Modos de IA

Scripts

npm run dev        # Servidor de desarrollo
npm run local      # Build + produccion en un comando
npm run init       # Inicializar base de datos
npm run init:seed  # Inicializar + datos demo
npm run seed       # Solo datos demo
npm run mcp        # Servidor MCP para Claude Desktop
npm run build      # Build de produccion
npm start          # Servidor de produccion
npm run lint       # Verificar codigo

Contribuir

1. Fork el repositorio
2. Crea tu branch (git checkout -b feature/mi-feature)
3. Haz commit de tus cambios (git commit -m 'Agregar feature')
4. Push a tu branch (git push origin feature/mi-feature)
5. Abre un Pull Request

English

What is Auto-CRM?

Auto-CRM is an open-source CRM that automatically customizes itself to your business. Open it with Claude Code, run /setup, and the CRM adapts to your industry, pipeline, and workflow. Everything runs on your computer — your data never leaves your machine.

Why Auto-CRM?

• 100% local — SQLite database on your computer. Zero external services. Your data stays yours.
• Self-customizing — Run /setup and Claude adapts the pipeline, lead sources, language, and theme to your business.
• AI included — Classifies leads, analyzes your pipeline, suggests next steps. No API key required.
• Free forever — Open source. No subscriptions. No limits on contacts or deals.
• Ready in 3 commands — Clone, install, run. Your CRM working in under 2 minutes.

Quick Start

git clone https://github.com/Hainrixz/auto-crm.git
cd auto-crm && npm install
npm run dev

Open http://localhost:3000 — your CRM is ready.

# Optional: load demo data to explore
npm run init:seed

Customize with Claude Code

Open the project with Claude Code and type:

/setup

The assistant asks about:

1. Business type and industry
2. Sales pipeline stages
3. Where your leads come from
4. Language and visual theme

And customizes everything automatically.

Features

Dashboard

Main panel with real-time KPIs: total contacts, active deals, pipeline value, hot leads. Pipeline charts and recent activity.

Kanban Pipeline

Visual drag & drop board. Drag deals between stages. Each card shows value, contact, and lead temperature.

Contact Management

Table with search, temperature filters (cold/warm/hot), and source filters. Lead score for each contact. Click for full detail with history.

Quick Actions

WhatsApp, call, and copy buttons directly on each contact. One click to open WhatsApp chat or start a call.

Activity Tracking

Timeline of all interactions: calls, emails, meetings, notes. Follow-up system with alerts for overdue items.

Lead Classification

Two modes: automatic rules (no API key) or AI with Claude (optional). Score from 0-100 and automatic temperature.

Webhook

Receive leads automatically from web forms (Typeform, Tally, Google Forms, Zapier). Supports fields in Spanish and English.

CSV Export

Download your contacts and deals as CSV with one click. Excel compatible.

Notifications

Dashboard banner for overdue follow-ups. Optional browser notifications every 5 minutes.

Email Digest

Daily email summary with pending follow-ups, hot leads, and metrics. Requires Resend (free tier).

Claude Code Commands

Integrations

Webhook — Receive leads automatically

curl -X POST http://localhost:3000/api/webhook \
  -H "Content-Type: application/json" \
  -d '{"name":"John","email":"[email protected]","phone":"555-1234","company":"Acme Inc"}'

Supports Spanish field names (nombre, correo, telefono, empresa) and nested Typeform format.

Export data

# Download contacts as CSV
curl http://localhost:3000/api/export?type=contacts -o contacts.csv

# Download deals as CSV
curl http://localhost:3000/api/export?type=deals -o deals.csv

MCP — Claude Desktop / Claude.ai

Connect your CRM directly to Claude Desktop:

{
  "mcpServers": {
    "auto-crm": {
      "command": "npx",
      "args": ["tsx", "/path/to/auto-crm/mcp/crm-server.ts"]
    }
  }
}

Now you can tell Claude: "Show me my hot leads" or "Add a new contact" from any chat.

Tech Stack

Deployment

Option 1 — Local development

npm run dev

Option 2 — Local production

npm run local

Option 3 — Docker

docker compose up -d

Option 4 — MCP (Claude Desktop)

Add the MCP config shown above to your claude_desktop_config.json.

Project Structure

auto-crm/
├── CLAUDE.md                # Instructions for Claude Code
├── .claude/commands/        # 8 interactive commands
├── mcp/crm-server.ts        # MCP server (10 tools)
├── scripts/init.ts          # Database initialization
├── src/
│   ├── app/                 # Pages and 13 API routes
│   ├── components/          # 28 React components + 21 shadcn/ui
│   ├── db/                  # SQLite schema + Drizzle ORM
│   ├── lib/                 # Utilities (scoring, AI, constants)
│   └── types/               # TypeScript types
├── data/crm.db              # Database (auto-generated)
├── Dockerfile               # Docker container
└── docker-compose.yml       # Docker Compose

Environment Variables

All are optional. The CRM works completely without any of them.

AI Modes

Scripts

npm run dev        # Development server
npm run local      # Build + production in one command
npm run init       # Initialize database
npm run init:seed  # Initialize + demo data
npm run seed       # Demo data only
npm run mcp        # MCP server for Claude Desktop
npm run build      # Production build
npm start          # Production server
npm run lint       # Check code

Contributing

1. Fork the repository
2. Create your branch (git checkout -b feature/my-feature)
3. Commit your changes (git commit -m 'Add feature')
4. Push to your branch (git push origin feature/my-feature)
5. Open a Pull Request