home / mcp / mcp-rag mcp server

MCP-RAG MCP Server

Provides a retrieval-augmented generation server integrating GroundX and OpenAI for document context with MCP.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "apatoliya-mcp-rag": {
      "command": "mcp",
      "args": [
        "dev",
        "server.py"
      ],
      "env": {
        "BUCKET_ID": "YOUR_BUCKET_ID",
        "OPENAI_API_KEY": "YOUR_OPENAI_API_KEY",
        "GROUNDX_API_KEY": "YOUR_GROUNDX_API_KEY"
      }
    }
  }
}

This MCP server provides a retrieval-augmented generation workflow that combines GroundX document retrieval with OpenAI models, all orchestrated through Modern Context Processing. You can ingest PDFs, perform semantic searches, and customize how results are generated and ranked to suit your context-heavy applications.

How to use

You will run the MCP server locally and connect to it with an MCP client. Ingest documents, then ask questions or run queries to retrieve relevant content and generate contextual responses. Use the standard search and document ingestion tools to build your knowledge base and tailor the completion behavior with a small, typed configuration.

Key workflows you can perform:

How to install

Prerequisites you need on your machine before installing and running the server.

Python 3.12 or higher
OpenAI API key
GroundX API key
MCP CLI tools

# 1) Clone the repository
git clone <repository-url>
cd mcp-rag

# 2) Create and activate a virtual environment
uv sync
source .venv/bin/activate  # On Windows, use `.venv\Scripts\activate

Configuration and environment

Create and configure your environment file to supply credentials and IDs required by the MCP server.

# 1) Copy the example environment file
cp .env.example .env

# 2) Populate your environment variables in .env
GROUNDX_API_KEY="your-groundx-api-key"
OPENAI_API_KEY="your-openai-api-key"
BUCKET_ID="your-bucket-id"

Available tools

ingest_documents

Ingests a document (for example a PDF) into the MCP server and processes it for later retrieval.

process_search_query

Performs a search against ingested documents and returns a structured response with query, score, and result.

SearchConfig

Configures how a search should be performed, including model selection and bucket use.