What is an MCP Server?

An MCP (Model Context Protocol) server provides AI assistants with access to local documents and information, enhancing their understanding and relevance.

How does an MCP Server work?

It serves document files from local source folders, classifies document types, and allows semantic search using RAG (Retrieval-Augmented Generation).

What are the key features of an MCP Server?

Key features include local document access, multiple source folder support, automatic document type classification, type-based filtering, semantic search with RAG, and offline functionality.

What types of documents can an MCP Server handle?

It supports Markdown, MDX, and text files organized as documentation, guides, or conventions.

How do I install an MCP Server?

You can install it using pip or directly from GitHub. RAG capabilities require the mcp-local-context[rag] installation.

How do I run the MCP Server?

Use the mcp-local-context command with options for specifying source directories, host, and port.

How do I configure the MCP Server for AI clients like Cursor or Claude Desktop?

Add a JSON configuration snippet to your MCP client’s configuration file, specifying the server URL or command.

What are the different ways to search for documents using an MCP Server?

You can list all documents, list documents by type, search by file path, or perform semantic searches based on content.

How do I add my own documents to the MCP Server?

Simply add your files to the configured source folders. The server automatically detects, classifies, and serves them.

What environment variables can be used to configure the MCP Server?

You can use SOURCE_DIRS, MCP_HOST, MCP_PORT, and MCP_PATH to configure the server.

Why use MCP Servers from UBOS Asset Marketplace?

UBOS platform simplifies finding and deploying MCP servers, enhances code generation, streamlines documentation retrieval, improves compliance, accelerates onboarding, and improves customer support.