Article - CS475917
Accessing a Windchill AI Assistant MCP Endpoint from a Python Client
Modified: 27-Jul-2026
Applies To
- Windchill PDMLink 13.0.2.11
- Windchill PDMLink 13.1.3.0
- Windchill PDMLink 12.1.2.23
Description
- This article describes an approach for integrating access to the Windchill MCP endpoint into a Python application or framework that uses the Anthropic MCP Python SDK.
- It shows how to complete the SAML login and then hand the resulting session to the SDK's HTTP client, after which the endpoint is used through the SDK in the usual way.
- The article documents the motivation, the Windchill authentication model, and the key implementation techniques; how the surrounding application obtains credentials, manages its lifecycle, and consumes the tools is outside its scope.
-
This article describes the recommended approach for connecting an MCP (Model Context Protocol) client to an SSO-protected Windchill environment when you are developing or embedding your own MCP client and have full control over the client implementation
- When to use this approach: This method is appropriate when you are building or embedding your own MCP client in code and therefore control the client.
- What It provides: The direct integration approach to perform the SSO login inside your own client and hand the authenticated session to your MCP SDK's HTTP client - no separate proxy process.
- Audience: Developers integrating MCP into their own application/framework.
- Important: PTC does not ship a ready-made MCP Client Code. The article provides illustrative code snippets only; the customer is responsible for building the direct SDK integration potentially using the below article's guidance and then distribute/install it internally.
- Depending on the client environment and deployment constraints, refer other approaches in Knowledge Base articles below:
This is a printer-friendly version of Article 475917 and may be out of date. For the latest version click CS475917