McpServer
A Model Context Protocol (MCP) server for supermarket inventory and sales management. Provides AI assistants with real-time access to product data, sales analytics, inventory tracking, and business intelligence tools through standardized MCP protocol integration.
claude mcp add --transport stdio darko-martinovic-mcpserver dotnet run --webapi \ --env ASPNETCORE_URLS="http://localhost:5000" \ --env ConnectionStrings__DefaultConnection="Server=your-server;Database=your-db;Integrated Security=true;TrustServerCertificate=true;"
How to use
This MCP server provides a dual-mode experience for a modern supermarket ecosystem. In Web API mode, it exposes versioned REST endpoints under /api/v1 for inventory, sales, revenue, and analytics via a plugin-based architecture that includes a SQL Server-backed Supermarket plugin and a MongoDB-backed ThirdApi plugin. The server also includes a unified Tool Proxy that lets you execute MCP tools by name through a single endpoint, and it supports Claude Desktop or MCP Inspector integrations in Console/MCP mode for AI-assisted workflows. In console/MCP mode, you can run the server as a tool provider without the web API surface, enabling Claude Desktop configurations to invoke tools locally. The API versioning strategy (e.g., /api/v1/...) ensures future compatibility as new endpoints and plugins are added. Tools can be executed via /api/v1/tool, probes via /api/v1/proxy/health, and data retrieval via the various /api/v1/supermarket and /api/v1/thirdapi endpoints. The system supports server-side logging, health checks, and plugin health endpoints for maintainability.
How to install
Prerequisites:
- .NET 9.0+
- SQL Server 2014+ (for Supermarket plugin)
- MongoDB 4.0+ (for ThirdApi plugin, optional)
Setup steps:
- Database: Create and seed the database
- Run Database/SetupDatabase.sql to create tables and sample data
- Configure: Edit appsettings.json with your SQL Server connection string
- Build:
dotnet build - Run (Web API mode):
dotnet run --webapi- This will start the REST API at http://localhost:5000 by default
- Run (Console/MCP mode):
dotnet run- This starts the server in console/MCP provider mode suitable for Claude Desktop or MCP Inspector integrations
- Optional Claude Desktop integration example (config):
{ "mcpServers": { "darko-martinovic-mcpserver": { "command": "dotnet", "args": ["run", "--project", "/path/to/your/McpServer", "--no-build"] } } }- Place this config where Claude Desktop expects it (see CLAUDE config guidance in README)
Notes:
- Ensure your appsettings.json contains a valid SQL Server connection string for the Supermarket plugin
- If you enable MongoDB interactions (ThirdApi), ensure MongoDB is running and the default connection string is configured
Additional notes
Tips and considerations:
- API Versioning: All endpoints are exposed under /api/v1 for forward compatibility; monitor api-supported-versions headers for available versions.
- Logging: Serilog-based logging writes to file and console; check Logs/mcpserver.log for issues.
- Environment variables: Override database connections via environment variables (e.g., ConnectionStrings__DefaultConnection) when deploying.
- Health checks: Use /health and /api/v1/proxy/health (tool proxy) to verify service health and plugin status.
- Claude Desktop integration: In Console/MCP mode, configure Claude to point to the MCP server and utilize the unified tool proxy endpoints to execute tools by name.
- Plugins: The architecture supports adding new plugins easily; each plugin provides tools, REST endpoints, and health checks.
Related MCP Servers
mcp-bigquery
A Model Context Protocol (MCP) server that provides secure, read-only access to BigQuery datasets. Enables Large Language Models (LLMs) to safely query and analyze data through a standardized interface.
awesome s
A curated list of excellent Model Context Protocol (MCP) servers.
Unity-AI-Animation
AI-powered tools for Unity animation workflow. Create and modify AnimationClips and AnimatorControllers directly through natural language commands.
penpot
MCP server for Penpot - Connect AI assistants to Penpot design platform via Model Context Protocol
mcp-azure-devops
Azure DevOps MCP for querying repositories, pipelines, work items, and delivery metadata via HTTP
mcpapp-colorpicker
A Model Context Protocol (MCP) App built with .NET that provides an interactive color picker with a rich UI