Build a ChatGPT Plugin Extension with Mastra

​ChatGPT plugins are moving beyond tools in a conversation. OpenAI’s new Plugin Extensions bring custom apps into key ChatGPT surfaces, including the sidebar, conversation panels, composer, and file viewers. Because these extensions build on the Model Context Protocol (MCP) Apps specification, developers can pair an interactive interface with agent tools and make the result feel like a native part of ChatGPT.

​In this workshop, we’ll build a ChatGPT extension with Mastra and trace the full path from a Mastra tool to an interactive MCP App running inside ChatGPT. You’ll see how the model-facing tool response and user-facing interface work together, how extension metadata controls where an app appears, and how the app communicates with its MCP server.

​We'll cover:

• ​Exposing Mastra tools and interactive resources through an MCP server

• ​Connecting a tool to an MCP App using ui:// resources, metadata, and structured content

• ​Adding an OpenAI extension entry point so the app can appear in ChatGPT surfaces such as the sidebar or conversation panel

• ​Testing the integration and understanding the communication flow between ChatGPT, the app interface, and Mastra tools

​Join us for a live demo with real code and the chance to ask questions directly to the engineers who built the feature.

​Hosted by

• ​Alex Booker, Developer Experience, Mastra

  • ​/bookercodes

  • ​https://mastra.ai

• ​Shane Thomas, Co-founder and CPO, Mastra

  • ​/smthomas3

  • ​/smthomas3

​Recording and code examples will be available to everyone who registers.