If you were one of the early adopters of OpenAI's tech and set out to build an AI assistant, you probably remember it being a bit of a coding odyssey. The chat completions API was efficient and capable, albeit very simple. It gets provided with a prompt, and it supplies a response. These were stateless calls, so it didn't remember conversations, which meant a lot of manual development in terms of maintaining chat states and integrating with various systems to augment the prompts with custom data.

The Azure OpenAI Assistants API provides a stateful API that simplifies the developer experience by keeping track of the conversation threads, freeing you from managing those endless chat histories or worrying about model prompt context limits. The Assistants API also provides a standard way of integrating external systems and custom logic via the code interpreter and function calling tools.

In this session, we will get up close with the Assistants API, learn all about its features, and find out how it streamlines the development of custom AI assistants.

You will learn:

• What the Assistants API brings to the table
• How to leverage tools with Assistants
• How to build custom tools