This guide shows how to use an Oracle Digital Assistant (ODA) LLM transformation handler with OCI Meta Llama 4 Maverick. The handler lets you shape and enrich prompts before they are sent to Maverick, giving you a central place to apply transformation logic such as adding contextual instructions, standardizing inputs, or preparing data for a consistent model response.
The accompanying sample demonstrates how to implement and configure the handler, then invoke Maverick from an ODA skill through the LLM Integration and Invoke Large Language Model component in Visual Flow Designer. It is intended to provide a practical starting point that you can adapt to your own conversational workflows, prompt standards, and business requirements.
This walkthrough assumes general familiarity with ODA development, including creating and configuring skills, using Visual Flow Designer, and working with LLM Integration and the Invoke Large Language Model component.
Note: you may have to click the above link and select “Save Link As…”