MCP Server
by donphi
The MCP Server enables AI assistants to query and retrieve information from custom document collections using the Model Context Protocol. This extends the AI's knowledge base with specific content, overcoming knowledge cutoffs.
Last updated: N/A
What is MCP Server?
The MCP Server is an implementation of the Model Context Protocol that allows AI assistants to access and retrieve information from external document collections. It processes documents, generates embeddings, and exposes the content through MCP tools.
How to use MCP Server?
To use the MCP Server, first clone the repository and configure the .env file. Place your documents in the data/ directory, then run the pipeline to process the files and generate embeddings. Finally, build and run the server, and configure your MCP-compatible AI assistant to connect to it using the generated mcp-config.json file.
Key features of MCP Server
Extends LLM knowledge with up-to-date information
Supports multiple embedding models (free and paid)
Provides a processing pipeline for documents
Exposes content through MCP tools
Use cases of MCP Server
Upgrading AI knowledge with latest framework documentation
Using private codebase documentation
Accessing technical specifications
Querying custom document collections
FAQ from MCP Server
What is the Model Context Protocol (MCP)?
What is the Model Context Protocol (MCP)?
The Model Context Protocol is a standard that enables AI assistants to access external tools and data sources.
What file types are supported?
What file types are supported?
By default, Markdown (.md), Text (.txt), PDF (.pdf), and Word documents (.docx, .doc) are supported. You can configure additional file extensions.
Do I need an OpenAI API key?
Do I need an OpenAI API key?
No, you can use free local embedding models without an API key. However, paid models require an OpenAI API key.
What if I get an 'Invalid reference format' error?
What if I get an 'Invalid reference format' error?
This usually means you haven't built the server image. Run docker-compose build server before running the server. Also, ensure you are using the latest version of the batch files from the repository.
How do I configure the server with my AI assistant?
How do I configure the server with my AI assistant?
Run the setup-mcpServer-json.sh (macOS/Linux) or setup-mcpServer-json.bat (Windows) script to generate a mcp-config.json file. Then, copy and paste the contents of this file into your AI assistant's MCP settings.