You’ve probably heard about OpenAI’s GPT3 or its sibling, ChatGPT, as they’ve been making headlines recently about their ability to generate very convincing human-like text. At this point, ChatGPT is only available in a web-based beta, but GPT3 is accessible via an API, and (great for TouchDesigner users!) has its own Python library. In this video, Jack DiLaura walks you through creating a simple, self-contained component for sending requests and receiving results from the GPT3 API. Along the way, you’ll learn more about the GPT3 API itself, as well as how to use DATs to quickly parse JSON data. Finally, you’ll take a look at the Engine COMP and how it can be very useful for avoiding performance hiccups by distributing processes like accessing an API.
Tutorial