Meal Planner Monorepo

A microservices-based meal planning application with AI-powered meal plan generation. Built with Go backend services, TypeScript frontend/agents, and containerized development environment.

Services Architecture

[React UI] ---> [API Gateway] ---> [Meal Service] ---> [PostgreSQL]
                      |                 |
                      v                 v
            [Agent Service] <---> [Logging Service]
                      |
                      v
              [MCP Service]

Core Services

• api-gateway/ – Go HTTP-to-gRPC proxy, main API entry point (port 8090)
• meal-service/ – Go gRPC service for meal/recipe data and business logic (port 50051)
• agent-service/ – TypeScript LangGraph AI agent for meal planning workflows (port 50053)
• mcp-service/ – Model Context Protocol server for agent interactions (port 3001)
• logging-service/ – Centralized Go gRPC logging service (port 50052)
• ui/ – React frontend application
• shared/ – Shared TypeScript types across services

Quick Start

# Install dependencies
yarn install

# Generate code (protobuf, OpenAPI, TypeScript clients)
make generate

# Start all services (PostgreSQL + all microservices)
docker-compose up -d

# Restore sample database
yarn db:restore

# Access the application
# UI: http://localhost:3000
# API: http://localhost:8090
# Swagger: http://localhost:8090/swagger/index.html

Development Commands

# Build all services
make build

# Run all tests
make test
yarn test                    # Aggregate test results

# Individual service tests
yarn test:backend           # Go services
yarn test:frontend          # React UI
yarn test:agent            # Agent service
yarn test:mcp              # MCP service

# Code generation
make generate              # All code generation
yarn generate_code         # TypeScript only

# Database operations
yarn db:backup             # Backup current database
yarn db:restore            # Restore from backup

# Formatting & linting
yarn format               # Format TypeScript + Go
make lint                 # Run all linters

# Individual service commands
yarn start:agent          # Run agent CLI
yarn start:mcp           # Run MCP server
yarn dev                 # Start UI in dev mode

Tech Stack

• Backend: Go 1.24, gRPC, PostgreSQL, Protocol Buffers
• Frontend: React 18, TypeScript 5.8, Yarn 4.9
• AI/Agent: LangGraph, OpenAI, Model Context Protocol
• Infrastructure: Docker Compose, Air (hot reload)
• Code Gen: protoc, Swagger/OpenAPI, Connect RPC

Requirements

• Go 1.22+
• Node.js 22+
• Docker & Docker Compose
• Protocol Buffers compiler (protoc)

Key Files

• CLAUDE.md – Project development guidelines
• Makefile – Build automation and code generation
• docker-compose.yml – Full development environment
• package.json – Yarn workspace configuration
• Each service has its own CLAUDE.md with specific dev notes