mcp -trueRAG
Model Context Protocol (MCP) server to access an instance of TrueRAG
claude mcp add --transport stdio ad-veritas-mcp-server-truerag uv --directory {path_to_mcp_server}/mcp-server-trueRAG run fastmcp run server.py \
--env GRAPHQL_API_KEY="your_api_key" \
--env GRAPHQL_ENDPOINT="your_graphql_endpoint"How to use
This MCP server provides a GraphQL policy API backed by the MCP Python SDK. It connects to a TrueRag GraphQL service to fetch and apply policy information through the MCP framework. You run the server using the uv tool and then interact with it via the MCP interface to retrieve, evaluate, or reason about policies in a consistent, protocol-driven way. The server exposes built-in capabilities for querying policy definitions, evaluating policy compliance, and integrating with clients that support the MCP protocol. To use it, ensure you have a valid GraphQL API key and endpoint, place them in a .env file, and start the server with uv as shown in the installation instructions. Once running, clients can connect to the MCP endpoint and issue standard MCP requests to fetch policy data, evaluate requests against policies, or reason about policy outcomes within your application.
How to install
Prerequisites:
- Python 3.8+ installed on your system
- uv tool installed (install via the provided install script in the README)
- Access to a TrueRag GraphQL API with a valid API key and endpoint
-
Clone the repository git clone https://github.com/Ad-Veritas/mcp-server-trueRAG.git cd mcp-server-trueRAG
-
Install uv (if not already installed)
macOS/Linux
curl -LsSf https://astral.sh/uv/install.sh | sh
Windows
powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"
-
Create a .env file in the repository root with GraphQL credentials GRAPHQL_API_KEY={your_api_key} GRAPHQL_ENDPOINT={your_graphql_endpoint}
-
Install Python dependencies (if a requirements file exists) python -m pip install -r requirements.txt
-
Run the MCP server using uv uv --directory {path_to_mcp_server}/mcp-server-trueRAG run fastmcp run server.py
-
Verify the server is running by checking logs or making a test MCP request to the server endpoint.
Additional notes
Tips and common considerations:
- Ensure the GraphQL API key and endpoint are kept secure and not committed to version control.
- The uv invocation requires the path to the mcp-server-trueRAG directory; adjust {path_to_mcp_server} accordingly.
- If you update Python dependencies, re-run installation steps as needed.
- Check for compatibility between the GraphQL API version and the MCP SDK version used by this server.
- For Claude or other MCP clients, point them to the MCP server and use the standard MCP request formats to retrieve policies, evaluate requests, or reason about policy outcomes.
Related MCP Servers
ai-engineering-hub
In-depth tutorials on LLMs, RAGs and real-world AI agent applications.
MAI-UI
MAI-UI: Real-World Centric Foundation GUI Agents ranging from 2B to 235B
5-Day-AI-Agents-Intensive-Course-with-Google
5-Day Gen AI Intensive Course with Google
sparql-llm
🦜✨ Chat system, MCP server, and reusable components to improve LLMs capabilities when generating SPARQL queries
pydata-london-2025
Multi-Agent AI Orchestration Workshop
jupyter
🪐 Model Context Protocol (MCP) Server for Jupyter