Summary

Cookbook update: "Building a Supply-Chain Copilot with OpenAI Agent SDK and Databricks MCP Servers"

• Migrated the Agent code repository from OpenAI’s GitHub to the official OpenAI Cookbook GitHub, following this example.
• Made the Databricks Supply Chain setup optional, clarifying that readers can either use their own Databricks datasets and analytical workloads or jump-start with the Databricks Supply Chain Solution Accelerator. The supporting repo was moved from a personal to my OpenAI GitHub account.
• In the integration section for Databricks MCP servers, added CLI-specific code so the Agent can be launched directly from the command line. Also included Databricks MCP authentication and Guardrails implementation examples.
• Enhanced the FastAPI section with clearer instructions and code to start the backend server, including guidance around the api_server.py file.
• In the React chat UI section, added frontend setup instructions and a code snippet demonstrating how the app streams the response token-by-token from the FastAPI /chat endpoint.
• Replaced the “Future Improvements” section with a more actionable “Next Steps” section tailored to readers.
• Performed several editorial and clarity-focused refinements across the following sections: Solution Overview, Architecture, Databricks Authentication Setup and Connecting to Databricks MCP Servers.
• Updated tags in the registry.yaml file.

Motivation

• This maintains consistency with the structure and style of existing cookbooks.