Skip to main content

​
MCP Server Integration Guide

Streamline your AI applications by connecting to the Praxis AI Middleware’s MCP (Model Context Protocol) server. The MCP server enables powerful features such as the search-instance-rag tool for in-depth context retrieval from digital twins. This guide walks you through configuration, connection, digital twin linking, and troubleshooting.

​
Overview

The Praxis AI Middleware includes an embedded MCP server, which exposes essential AI endpoints for seamless LLM integration. With tools like search-instance-rag, you can enhance your educational AI with real-time, context-aware reasoning—building robust, interconnected AI systems.
Key Benefits
  • Exposes your digital twin’s knowledge to your favorite LLM.
  • Integrates instances securely.
  • Powers sophisticated educational workflows.

​
Configure your Digital Twin as an MCP Server

Enable the MCP Server support and save your digital twin in the Configuration and Integrations panel MCP Server Enabled Switch

​
Connection Configuration

To connect your AI or middleware app to the MCP server:
  1. Server URL: https://pria.praxislxp.com/api/mcp
  2. Server Label: pria-mcp
  3. Description: (Optional) The description of your MCP server instructs the AI on when to use the tools configured. We suggest the following instruction:
Use this tool to run authoritative semantic search against the remote Digital Twin via the MCP server. Invoke search-instance-rag whenever the user’s question targets content stored in (prefer it over local search_rag for remote-scope queries). Input is a natural‑language query string; output is a relevance‑ordered list of matches containing each item’s file title, a concise summary, and a direct URL (include source_id and score when available). If no relevant results exist, return an empty list. Do not infer, synthesize, or modify content—return only what the remote instance provides.
  1. Authorization Header: Use Bearer authentication with your Digital Twin’s MCP Secret: 
    Authorization: Bearer <your-digital-twin-mcp-secret>
    Example Token:
    Bearer f831501f-b645-481a-9cbb-331509aaf8c1
Copy the token from the MCP Secret field from the Configuration and Integrations panel of your Digital Twin MCP Server Enabled Switch
  1. Primary Tool Available:
    • search-instance-rag (Contextual retrieval for digital twins)
    • api-admin-docs (Documents the Admin functions of the Praxis AI middleware REST API)
    • api-rt-docs (Documents the User-based Runtime functions of the Praxis AI middleware REST API)
    • Additional tools may be available depending on your configuration

​
Testing

You may use the MCP Inspector (see docs) or use any other tools to validate requests and see live responses and interact with your MCP Server. MCP Inspector
One practical use case for linking two digital twins via connector MCP is to enable knowledge sharing—where the document collection and contextual knowledge of one digital twin remain centralized, while other digital twins or systems can securely query and leverage this content1 . This approach avoids unnecessary data duplication, ensures that updates to source knowledge are instantly available to all connected twins, and allows different AI instances to collaborate or access real-time context without copying or synchronizing datasets. Ultimately, it streamlines data management, strengthens consistency, and unlocks powerful distributed workflows within connected AI ecosystems. To link two digital twins (e.g., “Twin B” connects with “Twin A”):
  1. In Digital Twin B:
    • Navigate to connectors.
    • Add a new connector with:
      • Type: Connector MCP
      • Target: Digital Twin A’s MCP Server endpoint (from Configuration)
      • Bearer token: Digital Twin A’s MCP Secret
  2. Set up connection parameters as shown below:
Connector MCP configuration to Digital Twin A.
The Bearer token you provide here should match Digital Twin A’s MCP Secret. This ensures secure, authenticated communication.
  1. Save and enable your connector.
Connector MCP List
  1. Launch the first request to test the connectivity
Test connectivity through your connector MCP
  1. And review the Agent Details
Review agwnt details

​
Troubleshooting & Tips

  • Invalid Token:
    Ensure your Authorization header matches an existing Digital Twin’s MCP Secret.
  • Connection Fails:
    Double-check endpoint URLs and enable connectors in both twins.
  • Not using the connector:
    Verify that the server label is configured correctly, and that the tools are named properly in the instance. Then ask a specific query like:
using the pria-mcp, execute the search-instance-rag tool to find information on ‘xyz’

​
More information

More information on Model Context Protocol available at MCP.io

Footnotes

  1. Model Context Protocol Documentation – Outlines knowledge sharing and centralized data strategies with MCP-connected digital twins. ↩