Overview
What is McpServer?
McpServer is a .NET implementation of the Model Context Protocol (MCP) that integrates Ollama AI models with a book management API. It provides HTTP endpoints to execute tools for book CRUD operations, AI model interaction, and system utilities, targeting developers and integrators who need a bridge between LLMs and structured data.
How to use McpServer?
Deploy using Docker Compose (recommended) or run locally with .NET 9. After startup, access the server at http://localhost:5500 and call tools via POST /api/execute. Integrate with Open WebUI by setting MCP_SERVER_URL environment variable.
Key features of McpServer?
- Book CRUD tools: create, read, update books via API.
- Ollama integration: list models and generate AI responses.
- Connectivity and configuration diagnostic utilities.
- HTTP API with OpenAPI documentation at
/api/openapi.json. - Docker Compose orchestration for single-command deployment.
- Works with Open WebUI for chat-based tool invocation.
Use cases of McpServer?
- AI-powered book search and retrieval using natural language.
- Automated book cataloging through LLM-generated prompts.
- Educational demonstration of MCP protocol with a live API.
- Combined AI response generation with real-time book database queries.
FAQ from McpServer
How to check connectivity between McpServer components?
Use the VerificarConectividade tool by sending a POST request to /api/execute with tool name VerificarConectividade and no parameters.
How to list available Ollama models through McpServer?
Call the ListModels tool via POST /api/execute with tool name ListModels and no parameters.
What are the system requirements for running McpServer?
At least 8 GB of RAM (16 GB recommended for larger AI models), 10 GB of disk space, and Docker installed. .NET 9 SDK is needed for local development.
How to add a new book using McpServer?
Use the CadastrarAsync tool with a JSON body containing titulo, autor, anoPublicacao, and genero fields, e.g., via POST /api/execute.
How to troubleshoot connectivity issues between services?
Ensure all Docker containers are running (docker-compose ps), check logs (docker-compose logs), and use the VerificarConectividade tool to verify component links.
