evolution-api/CLAUDE.md

CLAUDE.md

This file provides comprehensive guidance to Claude AI when working with the Evolution API codebase.

Project Overview

Evolution API is a powerful, production-ready REST API for WhatsApp communication that supports multiple WhatsApp providers:

• Baileys (WhatsApp Web) - Open-source WhatsApp Web client
• Meta Business API - Official WhatsApp Business API

Built with Node.js 20+, TypeScript 5+, and Express.js, it provides extensive integrations with chatbots, CRM systems, and messaging platforms in a multi-tenant architecture.

Common Development Commands

Build and Run

# Development
npm run dev:server    # Run in development with hot reload (tsx watch)

# Production
npm run build        # TypeScript check + tsup build
npm run start:prod   # Run production build

# Direct execution
npm start           # Run with tsx

Code Quality

npm run lint        # ESLint with auto-fix
npm run lint:check  # ESLint check only
npm run commit      # Interactive commit with commitizen

Database Management

# Set database provider first
export DATABASE_PROVIDER=postgresql  # or mysql

# Generate Prisma client (automatically uses DATABASE_PROVIDER env)
npm run db:generate

# Deploy migrations (production)
npm run db:deploy      # Unix/Mac
npm run db:deploy:win  # Windows

# Development migrations (with sync to provider folder)
npm run db:migrate:dev      # Unix/Mac
npm run db:migrate:dev:win  # Windows

# Open Prisma Studio
npm run db:studio

# Development migrations
npm run db:migrate:dev      # Unix/Mac
npm run db:migrate:dev:win  # Windows

Testing

npm test    # Run tests with watch mode

Architecture Overview

Core Structure

• Multi-tenant SaaS: Complete instance isolation with per-tenant authentication
• Multi-provider database: PostgreSQL and MySQL via Prisma ORM with provider-specific schemas and migrations
• WhatsApp integrations: Baileys and Meta Business API with unified interface
• Event-driven architecture: EventEmitter2 for internal events + WebSocket, RabbitMQ, SQS, NATS, Pusher for external events
• Microservices pattern: Modular integrations for chatbots, storage, and external services

Directory Layout

src/
├── api/
│   ├── controllers/     # HTTP route handlers (thin layer)
│   ├── services/        # Business logic (core functionality)
│   ├── repository/      # Data access layer (Prisma)
│   ├── dto/            # Data Transfer Objects (simple classes)
│   ├── guards/         # Authentication/authorization middleware
│   ├── integrations/   # External service integrations
│   │   ├── channel/    # WhatsApp providers (Baileys, Business API)
│   │   ├── chatbot/    # AI/Bot integrations (OpenAI, Dify, Typebot, Chatwoot)
│   │   ├── event/      # Event systems (WebSocket, RabbitMQ, SQS, NATS, Pusher)
│   │   └── storage/    # File storage (S3, MinIO)
│   ├── routes/         # Express route definitions (RouterBroker pattern)
│   └── types/          # TypeScript type definitions
├── config/             # Environment and app configuration
├── cache/             # Redis and local cache implementations
├── exceptions/        # Custom HTTP exception classes
├── utils/            # Shared utilities and helpers
└── validate/         # JSONSchema7 validation schemas

Key Integration Points

Channel Integrations (src/api/integrations/channel/):

• Baileys: WhatsApp Web client with QR code authentication
• Business API: Official Meta WhatsApp Business API
• Connection lifecycle management per instance with automatic reconnection

Chatbot Integrations (src/api/integrations/chatbot/):

• EvolutionBot: Native chatbot with trigger system
• Chatwoot: Customer service platform integration
• Typebot: Visual chatbot flow builder
• OpenAI: AI capabilities including GPT and Whisper (audio transcription)
• Dify: AI agent workflow platform
• Flowise: LangChain visual builder
• N8N: Workflow automation platform
• EvoAI: Custom AI integration

Event Integrations (src/api/integrations/event/):

• WebSocket: Real-time Socket.io connections
• RabbitMQ: Message queue for async processing
• Amazon SQS: Cloud-based message queuing
• NATS: High-performance messaging system
• Pusher: Real-time push notifications

Storage Integrations (src/api/integrations/storage/):

• AWS S3: Cloud object storage
• MinIO: Self-hosted S3-compatible storage
• Media file management and URL generation

Database Schema Management

• Separate schema files: postgresql-schema.prisma and mysql-schema.prisma
• Environment variable DATABASE_PROVIDER determines active database
• Migration folders are provider-specific and auto-selected during deployment

Authentication & Security

• API key-based authentication via apikey header (global or per-instance)
• Instance-specific tokens for WhatsApp connection authentication
• Guards system for route protection and authorization
• Input validation using JSONSchema7 with RouterBroker dataValidate
• Rate limiting and security middleware
• Webhook signature validation for external integrations

Important Implementation Details

WhatsApp Instance Management

• Each WhatsApp connection is an "instance" with unique name
• Instance data stored in database with connection state
• Session persistence in database or file system (configurable)
• Automatic reconnection handling with exponential backoff

Message Queue Architecture

• Supports RabbitMQ, Amazon SQS, and WebSocket for events
• Event types: message.received, message.sent, connection.update, etc.
• Configurable per instance which events to send

Media Handling

• Local storage or S3/Minio for media files
• Automatic media download from WhatsApp
• Media URL generation for external access
• Support for audio transcription via OpenAI

Multi-tenancy Support

• Instance isolation at database level
• Separate webhook configurations per instance
• Independent integration settings per instance

Environment Configuration

Key environment variables are defined in .env.example. The system uses a strongly-typed configuration system via src/config/env.config.ts.

Critical configurations:

• DATABASE_PROVIDER: postgresql or mysql
• DATABASE_CONNECTION_URI: Database connection string
• AUTHENTICATION_API_KEY: Global API authentication
• REDIS_ENABLED: Enable Redis cache
• RABBITMQ_ENABLED/SQS_ENABLED: Message queue options

Development Guidelines

The project follows comprehensive development standards defined in .cursor/rules/:

Core Principles

• Always respond in Portuguese (PT-BR) for user communication
• Follow established architecture patterns (Service Layer, RouterBroker, etc.)
• Robust error handling with retry logic and graceful degradation
• Multi-database compatibility (PostgreSQL and MySQL)
• Security-first approach with input validation and rate limiting
• Performance optimizations with Redis caching and connection pooling

Code Standards

• TypeScript strict mode with full type coverage
• JSONSchema7 for input validation (not class-validator)
• Conventional Commits enforced by commitlint
• ESLint + Prettier for code formatting
• Service Object pattern for business logic
• RouterBroker pattern for route handling with dataValidate

Architecture Patterns

• Multi-tenant isolation at database and instance level
• Event-driven communication with EventEmitter2
• Microservices integration pattern for external services
• Connection pooling and lifecycle management
• Caching strategy with Redis primary and Node-cache fallback

Testing Approach

Currently, the project has minimal formal testing infrastructure:

• Manual testing is the primary approach
• Integration testing in development environment
• No unit test suite currently implemented
• Test files can be placed in test/ directory as *.test.ts
• Run npm test for watch mode development testing

Recommended Testing Strategy

• Focus on critical business logic in services
• Mock external dependencies (WhatsApp APIs, databases)
• Integration tests for API endpoints
• Manual testing for WhatsApp connection flows

Deployment Considerations

• Docker support with Dockerfile and docker-compose.yaml
• Graceful shutdown handling for connections
• Health check endpoints for monitoring
• Sentry integration for error tracking
• Telemetry for usage analytics (non-sensitive data only)